paperID
stringlengths 36
36
| pwc_id
stringlengths 8
47
| arxiv_id
stringlengths 6
16
⌀ | nips_id
float64 | url_abs
stringlengths 18
329
| url_pdf
stringlengths 18
742
| title
stringlengths 8
325
| abstract
stringlengths 1
7.27k
⌀ | authors
stringlengths 2
7.06k
| published
stringlengths 10
10
⌀ | conference
stringlengths 12
47
⌀ | conference_url_abs
stringlengths 16
198
⌀ | conference_url_pdf
stringlengths 27
199
⌀ | proceeding
stringlengths 6
47
⌀ | taskID
stringlengths 7
1.44k
| areaID
stringclasses 688
values | embedding
stringlengths 9.26k
12.5k
| umap_embedding
stringlengths 29
44
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
aded5510-56ac-4e25-89d5-d6cfce0edf5f
|
exploring-the-relationship-between-center-and
| null | null |
http://dx.doi.org/10.1109/tcsvt.2022.3218284
|
http://dx.doi.org/10.1109/tcsvt.2022.3218284
|
Exploring the Relationship between Center and Neighborhoods: Central Vector oriented Self-Similarity Network for Hyperspectral Image Classification
|
To mine the spectral-spatial information of target pixel in hyperspectral image classification (HSIC), convolutional neural network (CNN)-based models widely adopt patch-based input pattern, where a patch represents its central pixel and the neighbor pixels play auxiliary roles in the classification process. However, compared to the central pixel, its neighbor pixels often have different contributions for classification. Although many existing patch-based CNNs could adaptively emphasize the spatial neighbor information, most of them ignore the latent relationship between the center pixel and its neighbor pixels. Moreover, efficient spectral-spatial feature extraction has been a difficult yet vital topic for HSIC. To address the mentioned problems, a central vector oriented self-similarity network (CVSSN) is proposed for HSIC. Specifically, based on two similarity measures, we firstly design an adaptive weight addition based spectral vector self-similarity module (AWA-SVSS) in input space and a Euclidean distance based feature vector self-similarity module (ED-FVSS) in feature space to fully mine the central vector oriented spatial relationships. Besides, a spectral-spatial information fusion module (SSIF) is formulated as a new pattern to fuse the central 1D spectral vector and the corresponding 3D patch for efficient spectral-spatial feature learning of the subsequent modules. Moreover, we implement a channel spatial separation convolution module (CSS-Conv) and a scale information complementary convolution module (SIC-Conv) for efficient spectral-spatial feature learning. Extensive experimental results on four popular HSI data sets demonstrate the effectiveness and efficiency of the proposed method compared with other state-of-the-art methods. The source code is available at https://github.com/lms-07/CVSSN.
|
['and Gongping Yang', 'Yuwen Huang', 'Guangkuo Xue', 'Yikun Liu', 'Mingsong Li']
|
2022-10-31
|
exploring-the-relationship-between-center-and-1
|
https://ieeexplore.ieee.org/document/9933425?arnumber=9933425&tag=1
|
https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9933425
|
ieee-transactions-on-circuits-and-systems-for-7
|
['hyperspectral-image-segmentation']
|
['computer-vision']
|
[ 5.19361496e-01 -5.62227130e-01 -5.00368662e-02 -2.52849281e-01
-2.94738382e-01 -2.55083770e-01 2.02932477e-01 1.29017800e-01
-3.75598818e-01 4.22306806e-01 -1.75913900e-01 -1.56760037e-01
-5.59235871e-01 -1.08352983e+00 -3.57385457e-01 -1.06885505e+00
1.67366937e-02 -4.31391567e-01 2.35377192e-01 -2.10957751e-01
2.80831426e-01 6.42110586e-01 -1.61878467e+00 1.53970927e-01
1.22094882e+00 1.48066008e+00 6.60934210e-01 2.49846682e-01
-2.43559390e-01 3.59839320e-01 -2.25437373e-01 3.83871108e-01
3.94785553e-01 -4.30164337e-01 -4.67618376e-01 2.43113890e-01
2.52314210e-01 4.18770090e-02 -3.81590694e-01 1.49451101e+00
4.16174233e-01 3.92017126e-01 6.57644331e-01 -1.15981221e+00
-6.98197424e-01 3.96924943e-01 -8.44629288e-01 1.54908881e-01
-2.66790569e-01 1.30645439e-01 6.71057522e-01 -1.00990701e+00
1.54518738e-01 6.50587261e-01 7.34715939e-01 -1.85239851e-01
-9.87588346e-01 -7.78913975e-01 7.87031502e-02 5.06834269e-01
-1.67359662e+00 -2.58710906e-02 1.22476292e+00 -2.62724787e-01
6.33557558e-01 3.11517626e-01 8.36152077e-01 3.40628028e-01
-1.33769959e-01 6.37837112e-01 1.11227620e+00 -2.89523900e-01
6.41040057e-02 -9.84638333e-02 1.18535221e-01 5.18254101e-01
7.15641887e-04 1.34151816e-01 -2.03629106e-01 1.74622193e-01
6.48478210e-01 4.41051900e-01 -7.80068517e-01 -2.01848805e-01
-1.05667353e+00 6.95139885e-01 1.16682601e+00 6.22455478e-01
-4.65780795e-01 -2.69005656e-01 7.73942983e-03 4.81841564e-02
2.73748517e-01 2.80357692e-02 -2.73029596e-01 5.36971331e-01
-1.00124538e+00 -4.60851714e-02 1.98206440e-01 6.56198144e-01
1.31820524e+00 8.10173303e-02 -4.03863005e-02 1.14821184e+00
1.94933861e-01 6.01939678e-01 4.79575396e-01 -8.12736928e-01
2.80707985e-01 8.94890964e-01 -2.78487861e-01 -1.39082730e+00
-5.69250643e-01 -9.15371239e-01 -1.43038511e+00 5.66869117e-02
-9.95590463e-02 5.20637352e-03 -7.26778626e-01 1.32678139e+00
2.65525520e-01 4.05938983e-01 1.61300451e-01 1.19680130e+00
9.83956397e-01 9.17706847e-01 -1.20893735e-02 -3.32869530e-01
1.10721278e+00 -7.12437212e-01 -3.49254817e-01 -1.18698187e-01
3.03624630e-01 -8.06828499e-01 6.53342783e-01 -8.81203562e-02
-6.69584870e-01 -8.10306311e-01 -1.15870178e+00 1.61706299e-01
-6.05691791e-01 5.36684752e-01 5.80889821e-01 1.17032431e-01
-7.11537719e-01 5.79940617e-01 -3.54381323e-01 -2.71542579e-01
5.76963484e-01 1.14308171e-01 -3.70823354e-01 -1.23826921e-01
-1.22811675e+00 3.78199250e-01 7.03789592e-01 4.00825381e-01
-1.62847564e-01 -7.54871726e-01 -8.35697770e-01 2.47697547e-01
2.47034207e-01 -2.65023589e-01 6.28434956e-01 -1.20330858e+00
-1.03624225e+00 3.95027131e-01 -1.46893114e-01 6.42502913e-03
-1.81448400e-01 5.21100163e-01 -8.29291821e-01 4.14099604e-01
3.19190085e-01 5.61096370e-01 7.66205072e-01 -1.19061768e+00
-9.70035970e-01 -3.87605518e-01 -2.51688838e-01 5.41797340e-01
-6.59107506e-01 -3.09704036e-01 -3.26410383e-01 -8.25536191e-01
6.82819247e-01 -5.17583549e-01 -9.28952992e-02 1.09469473e-01
-4.07090068e-01 -2.13557724e-02 1.08620644e+00 -4.90817189e-01
1.15770400e+00 -2.43429780e+00 -2.20657811e-02 4.85372573e-01
1.35196328e-01 5.19362748e-01 -3.47288460e-01 1.32346377e-01
-5.27990341e-01 -1.13332525e-01 -6.67408109e-01 4.23972398e-01
-4.05782223e-01 -2.16688961e-01 1.66735619e-01 5.89922130e-01
3.78647596e-01 7.20264077e-01 -8.21267366e-01 -3.60108167e-01
5.10857642e-01 5.64588010e-01 -6.78333119e-02 -1.01549782e-01
1.48838475e-01 1.65909886e-01 -3.69755536e-01 8.42897892e-01
1.28307056e+00 -2.67744511e-01 -2.07024828e-01 -8.44039321e-01
-6.93822742e-01 -4.43472385e-01 -1.28975153e+00 1.59147072e+00
-2.29848102e-01 4.04825181e-01 2.82134116e-01 -1.26431668e+00
1.10196567e+00 1.13597978e-02 6.82985783e-01 -7.15479970e-01
9.37751755e-02 4.05782819e-01 8.98590311e-02 -3.77001762e-01
3.14393163e-01 -6.36937916e-02 5.54755628e-01 -3.03832944e-02
9.72919837e-02 6.12069480e-02 -8.62725675e-02 -1.28020927e-01
6.54739320e-01 -1.78241894e-01 3.12851250e-01 -4.56859320e-01
1.06884170e+00 2.16575667e-01 7.85677314e-01 3.27963322e-01
-2.52879083e-01 6.08259857e-01 -4.65322435e-02 -3.67171854e-01
-8.09954464e-01 -6.34343326e-01 -4.75299895e-01 5.39110959e-01
5.60384095e-01 -1.18264303e-01 -4.76525903e-01 -3.72287095e-01
2.97431499e-02 3.63823265e-01 -4.72264975e-01 -1.59167752e-01
-1.83252424e-01 -8.75834286e-01 2.39578649e-01 4.27485794e-01
1.14873302e+00 -9.30876732e-01 -2.75880486e-01 2.79486954e-01
-1.31377742e-01 -6.39979899e-01 -4.52429116e-01 2.65165418e-01
-6.14705265e-01 -1.20213783e+00 -8.64140928e-01 -8.73318791e-01
5.67574203e-01 1.13959968e+00 3.98667932e-01 4.81124558e-02
-3.31600606e-01 2.19327658e-02 -4.56905484e-01 -2.57968068e-01
2.16352507e-01 3.64722713e-04 -4.14879262e-01 4.06664938e-01
3.93348664e-01 -7.90710211e-01 -9.12309527e-01 3.16335142e-01
-1.06038892e+00 3.40903342e-01 7.30342209e-01 9.95402575e-01
7.29727507e-01 6.77286744e-01 2.85404772e-01 -5.25916338e-01
2.25387797e-01 -5.74982882e-01 -5.84623337e-01 2.73028702e-01
-5.19729853e-01 -4.33537275e-01 8.18804383e-01 -2.31584802e-01
-9.77228165e-01 3.21167618e-01 -2.35628132e-02 -4.70454067e-01
-4.10990685e-01 8.95959318e-01 -3.38540763e-01 -3.58767539e-01
5.09139955e-01 8.67318690e-01 7.13843405e-02 -3.40251923e-01
2.80398354e-02 8.41184616e-01 7.05627501e-01 -2.36247197e-01
1.03171170e+00 4.10940826e-01 2.30156973e-01 -1.05892146e+00
-8.29858601e-01 -8.02389443e-01 -6.74484313e-01 -2.03217953e-01
8.69115829e-01 -1.11746347e+00 -6.90240443e-01 7.95968115e-01
-9.93883848e-01 -1.13178186e-01 -1.83025107e-01 4.89268005e-01
-4.44545187e-02 4.88648385e-01 -2.04532593e-01 -5.16160548e-01
-3.06013346e-01 -1.03390741e+00 9.68517363e-01 7.87360489e-01
5.73407292e-01 -8.94771814e-01 -3.05938601e-01 1.03667818e-01
6.18219197e-01 1.14755526e-01 8.64752412e-01 -2.38638133e-01
-4.54819471e-01 -6.40035644e-02 -8.24815989e-01 4.98180807e-01
3.65545064e-01 -1.71347056e-02 -8.65894675e-01 -1.51504785e-01
-7.15545863e-02 2.22529676e-02 1.16734147e+00 6.44786119e-01
1.50673354e+00 6.52575046e-02 -1.91349566e-01 1.13338542e+00
1.95845425e+00 4.53170031e-01 4.91814524e-01 4.21866298e-01
8.15083325e-01 4.55467731e-01 5.07049203e-01 3.98914039e-01
1.88799948e-01 3.00612092e-01 5.49629390e-01 -5.48741758e-01
4.67929244e-02 1.50286723e-02 -5.62791526e-02 5.70414186e-01
-2.98867553e-01 -6.42652735e-02 -5.96573770e-01 2.52967417e-01
-1.85529959e+00 -1.22676444e+00 -4.50114191e-01 1.90993571e+00
4.71229404e-01 -3.78677726e-01 -3.61022860e-01 4.62100774e-01
9.97342765e-01 4.38365132e-01 -7.42687583e-01 3.77090633e-01
-7.92467713e-01 1.09066568e-01 6.83184206e-01 1.50211811e-01
-1.45385695e+00 6.62878752e-01 4.02130556e+00 1.15030587e+00
-1.38254273e+00 6.78813383e-02 5.09304166e-01 2.84741193e-01
-1.54215321e-01 -1.07909115e-02 -3.53932768e-01 5.03204107e-01
1.83248490e-01 3.66129242e-02 5.94317019e-01 6.83854759e-01
2.62693703e-01 -2.54842281e-01 -2.67793238e-01 1.32680535e+00
2.58650649e-02 -1.48341918e+00 -1.41721562e-01 -4.02145647e-03
7.30819225e-01 6.85870722e-02 2.55123135e-02 -3.14269811e-02
-2.20198661e-01 -6.70072556e-01 5.67401946e-01 7.47686803e-01
9.52149212e-01 -8.56284738e-01 8.77477646e-01 2.01905534e-01
-1.81904602e+00 -3.58114243e-01 -7.82639027e-01 1.84250176e-01
-3.98929834e-01 8.21295559e-01 -1.01847155e-02 9.81154740e-01
8.05643559e-01 1.23812914e+00 -7.06413746e-01 1.14971733e+00
-7.09965527e-02 4.40820515e-01 -1.78111300e-01 2.34398261e-01
5.92553377e-01 -6.08651459e-01 4.63220417e-01 1.05982029e+00
7.23451793e-01 5.23398459e-01 2.17294142e-01 9.63278651e-01
1.40152037e-01 3.07146102e-01 -4.24566090e-01 6.67301118e-02
6.37499869e-01 1.70785391e+00 -6.80839062e-01 -2.28744060e-01
-5.31542540e-01 9.82385099e-01 -5.23734093e-02 5.11157751e-01
-5.45858860e-01 -8.42137575e-01 6.41263545e-01 -9.19030458e-02
3.59115720e-01 -5.07481545e-02 -1.19235814e-01 -9.67247427e-01
5.14397100e-02 -6.31702065e-01 3.34300607e-01 -9.84002113e-01
-1.24581492e+00 4.92793292e-01 -3.48794222e-01 -1.75251269e+00
6.56052530e-01 -7.06654310e-01 -9.51341093e-01 1.38821983e+00
-1.99566269e+00 -1.20558071e+00 -8.98320794e-01 9.33324993e-01
3.67299080e-01 -3.68280470e-01 6.98808432e-01 8.69832933e-02
-8.24948192e-01 1.56041354e-01 4.06677604e-01 2.57812619e-01
4.69527155e-01 -9.84858871e-01 -3.72879416e-01 9.70189273e-01
-3.47138345e-01 4.39694703e-01 -3.65572535e-02 -4.93113935e-01
-1.27878165e+00 -1.52153432e+00 3.75345945e-01 6.27931833e-01
6.21832967e-01 2.29488388e-01 -1.06172144e+00 1.55603848e-04
7.34452456e-02 3.35825324e-01 7.63514996e-01 -3.42690051e-01
-2.26464182e-01 -5.66884696e-01 -8.42582405e-01 2.59496897e-01
8.16393673e-01 -6.16970241e-01 -3.38461362e-02 4.18423176e-01
6.02874517e-01 -1.07308559e-01 -8.99081647e-01 8.07582200e-01
3.39728892e-01 -1.15589607e+00 9.57178831e-01 6.40058294e-02
3.69651109e-01 -9.16065514e-01 -4.34970468e-01 -1.42259896e+00
-8.60383570e-01 1.59409717e-01 4.23539251e-01 1.11541200e+00
1.95301324e-01 -6.53095901e-01 6.10041916e-01 -1.39273778e-01
-4.57269639e-01 -6.57612503e-01 -6.30285263e-01 -6.39288723e-01
-2.36702755e-01 -5.34372330e-01 7.90387154e-01 1.15283501e+00
-2.54141301e-01 2.15312108e-01 8.80115181e-02 7.20841348e-01
7.01081216e-01 5.81780732e-01 1.80423364e-01 -1.44640899e+00
6.69095516e-02 -8.37330580e-01 -4.14288700e-01 -6.08451843e-01
1.47674650e-01 -1.20614314e+00 -1.49470726e-02 -1.62234211e+00
2.76574314e-01 -5.16245067e-01 -6.58344030e-01 6.51581705e-01
-2.45508954e-01 3.26126099e-01 1.55366793e-01 4.66801047e-01
-1.64929897e-01 7.71366358e-01 1.28611791e+00 -5.92973530e-01
-2.49373823e-01 -7.65765756e-02 -7.24432349e-01 4.66202259e-01
9.95978355e-01 -5.23946099e-02 -3.73424768e-01 -1.62951291e-01
-1.49048865e-01 3.02912910e-02 6.76072299e-01 -1.43716753e+00
6.31130219e-01 -3.57309580e-01 8.85850012e-01 -8.67953539e-01
1.47793621e-01 -9.44497228e-01 1.69446841e-01 4.16591138e-01
-2.14779228e-02 -5.69143772e-01 1.14521787e-01 3.71364683e-01
-5.41929662e-01 -2.62500614e-01 9.76709783e-01 -1.64005294e-01
-1.04271257e+00 6.63830698e-01 -2.96925098e-01 -5.98276377e-01
8.73957038e-01 -4.22847003e-01 -4.81246471e-01 4.45118267e-03
-4.34675664e-01 2.46761099e-01 3.42884243e-01 7.11814314e-02
9.65523779e-01 -1.52913713e+00 -6.35143816e-01 5.17920554e-01
4.60749835e-01 1.35819584e-01 8.05897295e-01 9.77127552e-01
-4.80395138e-01 2.88873821e-01 -3.21752071e-01 -6.76104426e-01
-1.14062655e+00 3.39688271e-01 5.51888943e-01 3.68947238e-01
-3.56180191e-01 7.45601356e-01 2.13022724e-01 -4.46620196e-01
-1.63688689e-01 -1.03376567e-01 -4.06408906e-01 1.38431981e-01
5.14341176e-01 2.48684958e-01 1.69471037e-02 -9.24603105e-01
-2.75202602e-01 7.54002273e-01 4.19450969e-01 3.46761048e-01
1.52243960e+00 3.04681249e-02 -5.89818120e-01 8.45192224e-02
1.36142957e+00 -4.28339988e-01 -1.26787615e+00 -7.11787343e-01
-4.32684779e-01 -4.32357222e-01 6.26142502e-01 -7.19659150e-01
-1.37463295e+00 9.11672533e-01 9.07863975e-01 5.50087318e-02
1.72469735e+00 -3.68214577e-01 5.14384270e-01 3.38915855e-01
-5.38761355e-02 -1.08964837e+00 -2.65151381e-01 5.00492990e-01
7.19416201e-01 -1.32644629e+00 -2.01956257e-02 -5.66213608e-01
-3.14441085e-01 1.42793381e+00 6.62131429e-01 1.83472037e-01
1.05140007e+00 -6.96515217e-02 -1.11750863e-01 -2.43568018e-01
-7.85853043e-02 -5.55430293e-01 4.69487399e-01 6.78275049e-01
3.04002732e-01 2.07123175e-01 -3.57581191e-02 5.02550840e-01
1.06267184e-01 -3.26760888e-01 1.04338020e-01 6.86730385e-01
-7.00374484e-01 -6.56518519e-01 -5.30425787e-01 5.73945522e-01
1.98412016e-01 -3.09202820e-01 -5.34185022e-02 6.17824912e-01
7.12613523e-01 1.05724299e+00 1.79763705e-01 -7.72917211e-01
1.20801635e-01 -1.43785641e-01 -2.07512802e-03 -3.85574818e-01
-1.52197868e-01 4.44834292e-01 -6.36219978e-01 -3.76937836e-01
-8.13055933e-01 -6.14807069e-01 -1.33396816e+00 -1.49029389e-01
-3.85473549e-01 2.10866064e-01 6.45667374e-01 6.43186510e-01
2.43914887e-01 5.30424058e-01 1.06401360e+00 -1.02147365e+00
-3.98551747e-02 -9.66224790e-01 -1.08400774e+00 8.10327306e-02
1.55153215e-01 -5.23090303e-01 -3.23577195e-01 -4.08853255e-02]
|
[10.047791481018066, -1.6403043270111084]
|
9875d6c6-0589-4211-b87e-5986ccc172f8
|
no-gestures-left-behind-learning
| null | null |
https://aclanthology.org/2020.findings-emnlp.170
|
https://aclanthology.org/2020.findings-emnlp.170.pdf
|
No Gestures Left Behind: Learning Relationships between Spoken Language and Freeform Gestures
|
We study relationships between spoken language and co-speech gestures in context of two key challenges. First, distributions of text and gestures are inherently skewed making it important to model the long tail. Second, gesture predictions are made at a subword level, making it important to learn relationships between language and audio. We introduce Adversarial Importance Sampled Learning, which combines adversarial learning with importance sampling to strike a balance between precision and coverage. We substantiate the effectiveness of our approach through large-scale quantitative and user studies, which show that our proposed methodology significantly outperforms previous stateof-the-art approaches for gesture generation.
|
['Louis-Philippe Morency', 'Ryo Ishii', 'Dong Won Lee', 'Chaitanya Ahuja']
|
2020-10-01
| null | null | null |
findings-of-the-association-for-computational
|
['gesture-generation']
|
['robots']
|
[ 3.86497676e-01 7.66844749e-02 -3.61231565e-01 -4.30825680e-01
-1.28009880e+00 -7.61894226e-01 8.60137284e-01 -1.82076976e-01
-5.45987785e-01 7.47911751e-01 8.93993616e-01 -8.51881579e-02
1.33310691e-01 -5.27517378e-01 -8.62027228e-01 -4.58420873e-01
-2.04444140e-01 4.29623067e-01 9.67197865e-02 -2.66296506e-01
5.14933281e-02 3.82890463e-01 -1.23230541e+00 3.31793904e-01
3.90029222e-01 8.76321971e-01 -4.44322079e-01 1.03918481e+00
-1.75215706e-01 8.11661899e-01 -9.53964829e-01 -5.07677197e-01
1.79345414e-01 -4.60874200e-01 -6.14661574e-01 -2.92526484e-01
3.41098040e-01 -5.85971773e-01 -3.65723640e-01 9.02194142e-01
9.59573209e-01 8.62281248e-02 1.16859734e+00 -1.38628411e+00
-5.44274867e-01 1.09666348e+00 -3.97219986e-01 9.23801512e-02
6.17716730e-01 3.10795099e-01 1.31692958e+00 -6.47022605e-01
3.39545965e-01 1.53165090e+00 6.00011587e-01 8.31365645e-01
-1.13346374e+00 -1.03311479e+00 5.26450574e-01 -1.50505766e-01
-1.17144823e+00 -5.12358904e-01 6.70539498e-01 -3.43509078e-01
6.82491422e-01 3.93826336e-01 5.33220112e-01 1.88881218e+00
-1.01017982e-01 1.18766737e+00 9.50422227e-01 -4.58627671e-01
4.49851394e-01 -1.09498531e-01 -2.37491667e-01 1.36806190e-01
-2.02992901e-01 4.58234578e-01 -8.34863722e-01 -4.28837448e-01
7.92689681e-01 -2.84669369e-01 -6.00713566e-02 6.06904402e-02
-1.08794904e+00 7.29943514e-01 7.12660179e-02 1.62090704e-01
-3.69973779e-01 5.27586877e-01 3.12269330e-01 2.24987701e-01
3.15374255e-01 3.16768527e-01 -2.74341017e-01 -5.70094466e-01
-8.49311113e-01 5.73882461e-01 9.33741629e-01 1.01250470e+00
-9.82051790e-02 2.13614240e-01 -5.80210805e-01 7.12234855e-01
5.38891733e-01 7.89608836e-01 5.07029653e-01 -8.42886686e-01
4.84641463e-01 -1.42374158e-01 1.20806564e-02 -5.98068833e-01
-5.22821508e-02 9.95602757e-02 -8.03470016e-01 2.06773788e-01
7.95268297e-01 -5.22070706e-01 -8.40687394e-01 1.89086032e+00
1.52403191e-01 1.39361277e-01 -2.21127998e-02 8.41057658e-01
5.56872308e-01 3.32479149e-01 3.12725991e-01 -1.23304039e-01
1.06601393e+00 -5.98452687e-01 -7.92088568e-01 2.27032229e-02
-2.07879901e-01 -7.41300881e-01 1.50159061e+00 3.68445456e-01
-1.10617828e+00 -3.25521231e-01 -6.61788940e-01 2.62301683e-01
7.16008544e-02 -3.87885094e-01 6.40280664e-01 7.34736919e-01
-6.81886196e-01 4.37854707e-01 -7.87994564e-01 -7.92272314e-02
4.19867396e-01 3.23364377e-01 3.48186046e-02 2.78345048e-01
-1.31047690e+00 4.37689871e-01 9.38751623e-02 -3.74140590e-01
-7.69171774e-01 -7.21050322e-01 -6.82342112e-01 -9.59571972e-02
1.62202850e-01 -2.68009514e-01 1.69334471e+00 -9.53929186e-01
-1.86093032e+00 3.46939445e-01 -3.34071666e-02 -5.35806298e-01
8.75297427e-01 -5.74293911e-01 -2.85630614e-01 7.41740391e-02
-4.13466722e-01 6.70210481e-01 1.13361526e+00 -1.29407847e+00
-6.25243843e-01 -1.48223236e-01 1.00777134e-01 3.86754066e-01
-2.96195716e-01 6.22855537e-02 -3.30599040e-01 -1.39588511e+00
-3.50813806e-01 -8.32013428e-01 -1.11287072e-01 -1.80865638e-02
-4.64985162e-01 -1.07840545e-01 5.95356286e-01 -5.65648556e-01
1.14361751e+00 -1.93741906e+00 1.12868892e-02 4.39518392e-01
5.48142977e-02 6.03247471e-02 -3.00037086e-01 5.24077773e-01
2.03494057e-01 3.39716315e-01 1.51892006e-02 -2.71354139e-01
3.02726209e-01 1.74379408e-01 -8.31534445e-01 3.84930260e-02
1.26607746e-01 1.07303417e+00 -1.06075823e+00 -4.87242937e-01
5.86791188e-02 5.59830546e-01 -6.77106082e-01 5.23574233e-01
-5.19326568e-01 6.74614549e-01 -4.25686002e-01 7.81619668e-01
1.16594009e-01 2.49601305e-01 1.60035253e-01 5.83172999e-02
3.32337767e-01 4.16096956e-01 -1.07479882e+00 1.32948470e+00
-4.52343494e-01 5.70529222e-01 -6.02719262e-02 -6.55126214e-01
5.80679357e-01 5.18745661e-01 4.07166272e-01 -4.09085989e-01
2.19378754e-01 -3.83665375e-02 -3.92186195e-02 -4.21026736e-01
2.71048188e-01 -3.87324184e-01 -3.40024322e-01 8.28260958e-01
-1.16134867e-01 -5.62688828e-01 -3.35228384e-01 6.39408901e-02
1.01302183e+00 4.43740748e-02 4.40901905e-01 1.65526256e-01
2.37402841e-02 -4.30488169e-01 1.95209965e-01 1.04381108e+00
-3.13524455e-01 7.00549543e-01 6.25140488e-01 -1.01549588e-01
-7.90523112e-01 -1.35192263e+00 2.90957481e-01 1.48111403e+00
-1.35927826e-01 -5.48703922e-03 -8.52318823e-01 -9.39888120e-01
7.23998547e-02 7.61759341e-01 -7.37208545e-01 2.51657572e-02
-5.27154148e-01 -3.91792238e-01 1.04788470e+00 8.40205550e-01
-2.68242545e-02 -1.33676195e+00 -3.87979180e-01 8.22389647e-02
-1.56213924e-01 -1.03391659e+00 -9.02878940e-01 1.46056926e-02
-5.71353197e-01 -8.92449796e-01 -9.43149388e-01 -3.77776891e-01
2.14846343e-01 -2.91103154e-01 1.22785211e+00 -1.52984872e-01
-1.03138000e-01 6.41724408e-01 -7.06014931e-01 -7.16042697e-01
-7.17653573e-01 1.81093365e-01 1.90637663e-01 -9.03747510e-03
3.72596741e-01 -8.40243459e-01 -4.52199817e-01 1.00295275e-01
-1.05661392e+00 -4.23778653e-01 7.10482061e-01 1.00318241e+00
4.67714310e-01 -2.29724810e-01 6.48240626e-01 -7.74789691e-01
1.01905859e+00 -3.50064725e-01 -2.54076838e-01 2.35197365e-01
-2.82642245e-01 3.41222554e-01 3.51906121e-01 -1.11884868e+00
-7.01441467e-01 1.43304050e-01 -5.42004228e-01 -2.37556532e-01
-3.10315073e-01 9.43786874e-02 -2.29310244e-01 2.32998103e-01
5.24874985e-01 2.55751729e-01 6.55336678e-02 -2.14697808e-01
5.95493317e-01 9.54002976e-01 6.88636005e-01 -7.82341480e-01
9.26492393e-01 3.85800302e-01 -3.19147408e-01 -9.04689908e-01
-5.71860135e-01 -1.61355034e-01 -6.31850600e-01 -1.95435807e-01
6.83845818e-01 -6.39283061e-01 -8.18525732e-01 5.73349476e-01
-1.11608326e+00 -7.35093355e-01 -6.72069430e-01 4.18377161e-01
-7.79987812e-01 1.98332384e-01 -5.77359855e-01 -1.21567059e+00
-4.23707098e-01 -9.22654212e-01 1.39406347e+00 -6.02263957e-02
-8.23694170e-01 -8.61882567e-01 5.75216301e-02 5.51864058e-02
3.06867659e-01 2.68096417e-01 7.24163055e-01 -9.24418688e-01
-4.27365541e-01 -2.32043818e-01 7.52937719e-02 3.20166320e-01
2.09139213e-01 -1.71228815e-02 -1.06118846e+00 -1.49643794e-01
-4.04024690e-01 -7.12886453e-01 5.57092011e-01 3.38889092e-01
1.05692708e+00 -5.60586810e-01 5.42009175e-02 3.38951737e-01
6.92032397e-01 8.90275314e-02 4.59433913e-01 -6.38580620e-02
5.97442389e-01 6.58970237e-01 7.55380809e-01 6.79628670e-01
3.46027583e-01 7.03024983e-01 2.97583789e-01 1.94020122e-01
-2.10829854e-01 -6.66130900e-01 4.47780341e-01 6.08761549e-01
1.93159804e-02 -3.18183780e-01 -8.70130956e-01 5.54835856e-01
-1.64925277e+00 -1.07510591e+00 4.63094592e-01 2.14027691e+00
1.36583304e+00 3.02242041e-01 5.75080514e-01 3.17412198e-01
2.09260821e-01 2.31768966e-01 -5.02574503e-01 -1.93016738e-01
2.23042086e-01 4.06905025e-01 4.73501325e-01 7.90600955e-01
-1.17855155e+00 9.60648715e-01 7.69237995e+00 9.19341207e-01
-1.12654161e+00 6.28836378e-02 5.34944117e-01 -3.57801229e-01
-5.38965285e-01 -5.92612386e-01 -7.24954188e-01 5.30840218e-01
6.19291365e-01 1.32291357e-03 5.90363264e-01 6.26286387e-01
3.37818898e-02 3.84440541e-01 -1.24010062e+00 6.83600247e-01
-1.01750240e-01 -9.09904718e-01 2.89457142e-01 1.01749273e-02
6.68505073e-01 -1.41519263e-01 2.76397198e-01 4.19772178e-01
8.32693458e-01 -1.33682811e+00 9.84143794e-01 3.38905960e-01
1.09550798e+00 -6.86221957e-01 5.50962865e-01 4.17910904e-01
-1.08051324e+00 1.45971533e-02 2.33431786e-01 -1.29562438e-01
2.88564235e-01 1.62174895e-01 -9.89106059e-01 -1.27570063e-01
5.19964278e-01 8.48518237e-02 8.05633441e-02 4.16141629e-01
-5.44737756e-01 8.89471710e-01 -4.56702292e-01 -3.79972726e-01
-2.08490714e-02 5.20204902e-01 6.45004511e-01 1.43711042e+00
9.52301919e-02 1.73657849e-01 1.83472484e-01 7.19257355e-01
-2.06648901e-01 6.73182458e-02 -6.43432796e-01 -2.97318667e-01
9.42599475e-01 5.50349832e-01 -4.47815448e-01 -2.34177500e-01
-2.36087769e-01 9.58818316e-01 3.03904600e-02 4.10156757e-01
-6.81418896e-01 -1.37473449e-01 1.02141690e+00 -7.40004852e-02
3.27083409e-01 -2.11633384e-01 -4.76799190e-01 -9.89439011e-01
-1.03810683e-01 -1.33358145e+00 2.34878987e-01 -1.42883852e-01
-1.51004148e+00 5.81491709e-01 8.22698548e-02 -1.25401986e+00
-9.58439112e-01 -2.98574179e-01 -4.92567480e-01 9.10902917e-01
-1.27317560e+00 -1.33128190e+00 -4.90444750e-02 4.89885092e-01
8.68021429e-01 -2.78224140e-01 1.02123368e+00 1.30505785e-02
-3.57073508e-02 1.30933797e+00 -1.36570603e-01 3.96997631e-01
6.73698306e-01 -1.33610117e+00 9.22760189e-01 5.19367337e-01
5.86700261e-01 4.79315609e-01 7.20921695e-01 -6.90925241e-01
-1.21197224e+00 -9.56719935e-01 7.24437475e-01 -7.04047024e-01
6.56565368e-01 -6.35463834e-01 -5.58133662e-01 4.46578026e-01
-1.48950636e-01 -9.49099287e-02 8.47286522e-01 1.15511633e-01
-6.34666800e-01 2.16381967e-01 -1.16919398e+00 8.43374431e-01
1.05304468e+00 -7.78755367e-01 -7.04879761e-01 5.72425500e-02
9.55456614e-01 -4.71593767e-01 -7.43629217e-01 3.73263866e-01
1.25250804e+00 -6.57964528e-01 1.12774599e+00 -6.86545789e-01
5.20129025e-01 3.19238901e-01 -4.92555976e-01 -1.45039809e+00
3.03445309e-01 -9.52622473e-01 -5.91957450e-01 1.32604635e+00
4.22517061e-01 -3.80318284e-01 8.56229365e-01 6.46860361e-01
5.70603013e-01 -6.25283301e-01 -7.72913396e-01 -7.73337066e-01
2.84644365e-01 -8.82421553e-01 8.52144361e-01 6.55716658e-01
-2.53290068e-02 4.01944548e-01 -6.17114186e-01 1.33074418e-01
5.98297536e-01 -2.70654500e-01 9.17220652e-01 -9.22571361e-01
-8.04291368e-01 -6.37650609e-01 -1.78542003e-01 -1.18566942e+00
1.96822554e-01 -3.20573777e-01 4.27144825e-01 -9.39944327e-01
-1.08845316e-01 -4.72572714e-01 -2.68171936e-01 3.33471805e-01
-3.64501655e-01 5.10672688e-01 2.32348159e-01 1.52009711e-01
-3.26511115e-01 6.40225828e-01 1.01036787e+00 -2.22205162e-01
-4.32323575e-01 5.47452152e-01 -7.15077162e-01 8.42882752e-01
5.92077613e-01 -3.67199183e-01 -5.09337544e-01 -2.45919511e-01
-4.03804146e-02 4.10368815e-02 1.37194559e-01 -7.85964429e-01
-7.68712163e-02 -3.61141890e-01 6.00618683e-02 -3.86310786e-01
3.91727418e-01 -8.86024714e-01 -4.96444702e-01 3.71017307e-01
-1.02590191e+00 -1.86504275e-01 8.82269070e-02 8.82460833e-01
-2.43348479e-01 1.15311697e-01 5.08819401e-01 1.09498501e-01
-2.80282795e-01 1.92633152e-01 -3.82273883e-01 5.03708243e-01
6.79996908e-01 6.04170337e-02 2.87041634e-01 -1.19737673e+00
-4.99002516e-01 9.67348069e-02 1.03920912e-02 6.83584392e-01
6.83462679e-01 -1.41038048e+00 -7.68805742e-01 3.51305127e-01
5.48927076e-02 -1.17877282e-01 -2.18450651e-01 3.77526194e-01
-2.46002689e-01 9.78656113e-02 3.56917940e-02 -3.27214867e-01
-1.48949265e+00 2.31136680e-01 -6.79685548e-02 -3.14300448e-01
-3.69424760e-01 1.05156648e+00 2.75421534e-02 -5.28166711e-01
1.04039180e+00 -5.51461577e-01 -7.92782679e-02 -1.49911446e-02
8.62996936e-01 1.88096628e-01 -4.59245563e-01 -5.04469693e-01
-1.89275250e-01 2.46317253e-01 1.14271224e-01 -8.50539267e-01
1.17141652e+00 1.80814549e-01 5.11772156e-01 7.49096572e-01
8.89834106e-01 3.65224898e-01 -1.53632236e+00 -2.80976325e-01
-1.27670631e-01 -3.70276421e-01 -2.79200763e-01 -9.77583826e-01
-6.79646671e-01 9.56181288e-01 5.90266347e-01 1.22939855e-01
1.11174357e+00 7.13403374e-02 1.05173588e+00 3.47493708e-01
3.50573212e-01 -9.15719867e-01 4.47392821e-01 6.74712121e-01
9.82674539e-01 -1.28386688e+00 -2.26057649e-01 -2.57945955e-01
-9.09995914e-01 7.58845568e-01 3.11803937e-01 -1.54703572e-01
6.92521513e-01 8.43163908e-01 4.16217893e-01 2.92405814e-01
-4.88620132e-01 -2.71761179e-01 5.78771710e-01 9.51927781e-01
6.23010457e-01 3.25963974e-01 -6.93201125e-02 8.55091810e-01
-5.41972101e-01 -6.22101203e-02 -4.86357957e-02 9.06128585e-01
-5.39739057e-02 -1.53416681e+00 -4.87348616e-01 3.64034146e-01
-8.07901919e-01 -2.70638824e-01 -7.77404487e-01 8.23355794e-01
-3.29518229e-01 7.60840297e-01 -2.73208208e-02 -6.97891831e-01
4.13734257e-01 3.18800211e-01 5.16539097e-01 -5.70611656e-01
-5.88240743e-01 2.20618490e-02 9.83763635e-02 -4.18737650e-01
-3.09158713e-01 -7.36032844e-01 -1.15528727e+00 -1.53686211e-01
-5.94744161e-02 -1.28611699e-01 5.28984547e-01 1.04834032e+00
5.83132543e-02 5.33929408e-01 5.33147037e-01 -1.11774218e+00
-1.29367971e+00 -1.23202634e+00 -5.84877372e-01 5.20965755e-01
5.94882250e-01 -3.95264477e-01 -2.24446490e-01 1.53985754e-01]
|
[5.629237174987793, -0.1108575388789177]
|
3f457c05-c7ae-45a3-999b-a7a89d1b43a6
|
deep-fast-vision-accelerated-deep-transfer
| null | null |
https://doi.org/10.5281/zenodo.7865289
|
https://doi.org/10.5281/zenodo.7865289
|
Deep Fast Vision: Accelerated Deep Transfer Learning Vision Prototyping and Beyond
|
Deep Fast Vision is a versatile Python library for rapid prototyping of deep transfer learning vision models. It caters to users of various levels, offering different levels of abstraction from high-level configurations for beginners to mid and low-level customization for professional data scientists and developers. Built around Keras and TensorFlow, this library also includes handy utilities.
|
['Fabi Prezja']
|
2023-04-26
| null | null | null |
zenodo-github-2023-4
|
['automl']
|
['methodology']
|
[-9.34719980e-01 -4.33663189e-01 -1.75718442e-01 -4.82849002e-01
-2.17617273e-01 -6.98438168e-01 5.19892573e-01 -1.59664020e-01
-5.43178260e-01 1.52132198e-01 -2.03654110e-01 -6.81092501e-01
2.79333085e-01 -4.86132592e-01 -2.19169721e-01 -3.81795913e-01
-2.63788730e-01 3.08927476e-01 5.47602177e-01 -4.11131717e-02
6.14099354e-02 7.08503723e-01 -1.52619886e+00 5.58670282e-01
2.14556366e-01 8.63392949e-01 2.57470042e-01 1.19717574e+00
-1.14788800e-01 1.05646014e+00 -4.47420299e-01 -4.11478698e-01
4.84028637e-01 3.04424852e-01 -8.98023844e-01 -2.87860721e-01
5.22324324e-01 -6.14797533e-01 -3.85163367e-01 7.62341857e-01
5.14669657e-01 -2.01793835e-01 1.35496795e-01 -1.57473040e+00
-7.57332742e-01 -7.02277292e-03 -3.92381966e-01 8.83509994e-01
1.85944512e-01 7.19078898e-01 9.63482797e-01 -1.17741144e+00
2.83046603e-01 1.23082185e+00 1.03233886e+00 4.87507641e-01
-1.05244088e+00 -7.29428232e-01 8.13271254e-02 3.17521095e-01
-9.13854003e-01 -5.89058280e-01 3.39305133e-01 -1.00580978e+00
1.37568951e+00 -2.08122078e-02 6.22299433e-01 1.26424730e+00
1.20051555e-01 9.29168284e-01 1.03310943e+00 -1.08440205e-01
2.23469645e-01 2.22166970e-01 7.30633378e-01 7.14093328e-01
-2.41834223e-02 -9.66727510e-02 -5.18095255e-01 -5.03784478e-01
1.16451466e+00 1.68397203e-01 1.02345526e-01 -2.12698448e-02
-1.29123306e+00 8.87314260e-01 4.56158996e-01 1.10324793e-01
-5.89179754e-01 8.50734338e-02 6.72631979e-01 4.39852268e-01
2.66371280e-01 1.15194336e-01 -8.45618784e-01 -3.99427384e-01
-7.29867280e-01 3.81454080e-01 8.33509862e-01 1.18114173e+00
8.02901387e-01 3.52488995e-01 -2.01918930e-02 8.28445375e-01
4.74085659e-01 -2.15787336e-01 5.71317673e-01 -1.06440568e+00
1.02671853e-03 2.56811380e-01 -9.43797156e-02 -3.95580918e-01
-4.36973453e-01 -4.60514694e-01 -6.08699620e-01 8.68495703e-01
3.06109428e-01 -2.48840705e-01 -1.03186011e+00 1.00963509e+00
1.60863608e-01 1.30145222e-01 -1.11709684e-01 7.41506636e-01
1.39921010e+00 5.56396782e-01 4.22905594e-01 4.11970645e-01
1.27507937e+00 -1.26383829e+00 -1.43465459e-01 -4.08723086e-01
1.39320180e-01 -9.58823740e-01 1.27855277e+00 4.42317069e-01
-1.06813765e+00 -7.67411351e-01 -8.82273018e-01 -4.69330162e-01
-5.10914147e-01 -5.14744222e-02 1.13104856e+00 4.99065280e-01
-1.57184505e+00 6.49589002e-01 -1.24311030e+00 -5.47226787e-01
6.17633939e-01 3.54418874e-01 -3.11914116e-01 2.46870029e-03
-7.40242720e-01 8.87001276e-01 2.68301398e-01 -7.69435912e-02
-9.19395328e-01 -1.07271612e+00 -6.71269417e-01 -4.32556905e-02
-3.49334151e-01 -1.03938568e+00 1.79635668e+00 -7.40124524e-01
-1.34606516e+00 9.47851300e-01 2.75400691e-02 -3.70644301e-01
5.12180805e-01 -1.56167699e-02 -2.21594796e-01 -6.76070107e-03
-1.04297348e-03 5.06993830e-01 8.79392624e-01 -6.93767667e-01
-6.30411625e-01 -2.46962279e-01 2.24680841e-01 9.69523117e-02
-2.08327189e-01 7.07530379e-01 -5.77656448e-01 -3.12610090e-01
-3.78646374e-01 -6.66101992e-01 -1.11949936e-01 7.28117108e-01
-2.11449601e-02 -2.29863316e-01 1.23728895e+00 -6.75486147e-01
3.06781143e-01 -2.09458280e+00 -3.12955052e-01 -7.29488507e-02
5.30067205e-01 4.84894335e-01 1.45317659e-01 3.79804790e-01
-2.67933279e-01 -2.78809279e-01 3.23846191e-01 -5.06150961e-01
-4.62310910e-02 3.02991331e-01 4.75316681e-02 2.62201846e-01
1.14293031e-01 7.78373599e-01 -9.14820015e-01 -2.79059529e-01
5.90928257e-01 6.54693604e-01 -5.26326716e-01 5.64812958e-01
1.19248167e-01 3.65702420e-01 -2.43261307e-01 9.88783121e-01
6.69551253e-01 -7.14516878e-01 -3.13452303e-01 1.84945501e-02
-4.24033165e-01 3.17252368e-01 -9.60141301e-01 1.64118707e+00
-4.78484571e-01 1.04123855e+00 6.58319652e-01 -5.92781782e-01
5.85359752e-01 2.02437609e-01 1.56555980e-01 -1.50080204e-01
-1.08225659e-01 -8.41634572e-02 -1.62027836e-01 -6.76847279e-01
3.95005435e-01 -1.02640010e-01 4.43705827e-01 2.60416448e-01
5.99711120e-01 -6.69931062e-03 2.62070298e-02 7.27965608e-02
1.16080260e+00 1.86358809e-01 1.96766436e-01 -3.39649260e-01
1.59471154e-01 -2.46703625e-02 4.07969296e-01 5.47347844e-01
-3.52735341e-01 4.00395006e-01 3.91771436e-01 -7.49241114e-01
-1.16839015e+00 -1.51379311e+00 -1.70897111e-01 1.82172441e+00
-7.04020619e-01 -5.47655106e-01 -6.03676081e-01 -2.37838954e-01
2.56245472e-02 1.29409522e-01 -4.56699103e-01 3.63240600e-01
2.62152366e-02 -5.11882663e-01 3.99341166e-01 7.72758186e-01
8.09328258e-01 -1.23009562e+00 -5.89721322e-01 -1.97963357e-01
2.52091825e-01 -1.56296468e+00 -1.66024417e-01 1.40841603e-01
-9.73824263e-01 -9.96159494e-01 -6.29138529e-01 -1.00034285e+00
3.60177875e-01 2.93170035e-01 1.45608866e+00 2.80981325e-02
-7.31613994e-01 6.29045844e-01 -1.20397069e-01 -6.53936923e-01
-2.61875749e-01 -1.74455479e-01 1.43943587e-03 -4.29655820e-01
4.86595720e-01 -5.68614006e-01 -6.76608443e-01 4.13752012e-02
-5.31566799e-01 2.07633704e-01 3.92127365e-01 5.61917782e-01
1.53693408e-01 -3.51874322e-01 4.08942848e-02 -4.04422969e-01
6.92418516e-01 -5.35646439e-01 -7.81078815e-01 -1.50406644e-01
-4.60823387e-01 -4.18022454e-01 6.48941576e-01 -1.47782132e-01
-8.21204185e-01 8.74061212e-02 -5.24542212e-01 -6.03562057e-01
-4.10537720e-01 4.17714536e-01 2.67766982e-01 -4.25045550e-01
8.98868740e-01 2.06109777e-01 1.10226758e-02 -8.76621306e-01
2.06573427e-01 8.87317836e-01 7.48936713e-01 -6.38553441e-01
5.88292599e-01 2.52364069e-01 -6.83559954e-01 -1.14169919e+00
-5.46590507e-01 -4.81382072e-01 -8.38150144e-01 -1.35090828e-01
9.14380312e-01 -1.10224366e+00 -8.13911498e-01 1.05415440e+00
-1.02990699e+00 -6.23950303e-01 -2.82185003e-02 3.94289762e-01
-1.99356675e-01 3.57943699e-02 -1.02711260e+00 -1.66112423e-01
-6.23216450e-01 -1.22079837e+00 6.37691975e-01 3.69870245e-01
-1.18183739e-01 -1.24821138e+00 7.21549317e-02 2.05619290e-01
6.00662589e-01 2.52327561e-01 4.76751029e-01 -4.68395591e-01
-4.81036991e-01 -2.47659951e-01 -5.84243655e-01 7.03152895e-01
7.25210235e-02 6.56550884e-01 -1.38965058e+00 -4.57693547e-01
-4.58210140e-01 -4.83737677e-01 4.94729251e-01 3.13626707e-01
1.35141468e+00 -1.47951484e-01 -1.62692085e-01 9.13036525e-01
1.21110237e+00 -5.35327531e-02 4.04838264e-01 5.89747965e-01
8.06734145e-01 2.46704176e-01 -1.35764793e-01 5.09145021e-01
6.34630919e-01 2.31639788e-01 4.81259644e-01 -5.20363927e-01
6.64145977e-04 1.77448004e-01 -2.03057192e-03 5.37698030e-01
-4.70395744e-01 8.04711044e-01 -1.24653327e+00 4.64786887e-01
-1.72948921e+00 -1.00471127e+00 -3.94249022e-01 2.01328611e+00
8.23443294e-01 2.97055542e-01 5.91103971e-01 -1.58543393e-01
4.49204028e-01 -4.16391976e-02 -4.66445595e-01 -8.03508103e-01
4.55040306e-01 4.43829417e-01 2.41197273e-01 1.32686436e-01
-1.05312634e+00 9.22230840e-01 8.54500866e+00 3.60986829e-01
-1.31040704e+00 4.08035815e-01 1.07177600e-01 -6.27797246e-02
1.98966399e-01 -6.88001886e-02 -7.18484044e-01 3.97034228e-01
1.02007031e+00 -3.93389910e-01 4.95590687e-01 1.37936068e+00
1.25830740e-01 1.00504912e-01 -1.18660021e+00 1.10601246e+00
-4.84992862e-01 -1.62453210e+00 -3.29845071e-01 -5.88829294e-02
3.01182240e-01 1.12959969e+00 -3.76725532e-02 6.60055459e-01
5.81629276e-01 -8.60598922e-01 5.89391530e-01 2.20302954e-01
5.82393408e-01 -5.56269407e-01 4.73828048e-01 1.83264032e-01
-9.79465961e-01 -1.13142043e-01 -5.34348607e-01 -6.22123957e-01
-3.19143683e-01 3.41404170e-01 -9.88945127e-01 -1.66361816e-02
1.34893131e+00 6.12391174e-01 -6.89750433e-01 1.28487027e+00
6.98112398e-02 6.09574080e-01 -2.68910706e-01 2.85846353e-01
3.46430272e-01 1.87438279e-01 -2.12159380e-02 1.91799927e+00
-1.59629285e-01 -2.40256503e-01 1.73352867e-01 6.54938042e-01
1.98023736e-01 -2.40445703e-01 -5.64045548e-01 1.06677815e-01
3.46015871e-01 1.67416716e+00 -4.50953335e-01 -4.05854225e-01
-8.95307362e-01 1.01849043e+00 5.88083863e-01 4.25082386e-01
-6.23774230e-01 -4.97512251e-01 1.55426335e+00 1.04852896e-02
3.54213834e-01 -8.89046013e-01 -3.47779572e-01 -9.75207806e-01
-3.70433748e-01 -7.13352025e-01 5.88795364e-01 -1.11845946e+00
-1.34191120e+00 6.31141603e-01 -4.62482050e-02 -1.00063920e+00
-1.30604699e-01 -1.21866226e+00 -1.06289101e+00 1.12057710e+00
-1.43103838e+00 -1.44227362e+00 -7.86444545e-01 9.72577870e-01
4.93349165e-01 -3.73414904e-01 1.23072267e+00 2.85099506e-01
-7.72695363e-01 3.69900823e-01 -2.47662067e-02 5.79413176e-01
5.25260806e-01 -1.38975430e+00 8.59351635e-01 5.93240619e-01
-2.74109364e-01 1.01711559e+00 7.61566520e-01 1.07718237e-01
-1.30933845e+00 -1.08733046e+00 5.34620106e-01 -3.87319922e-01
9.82343078e-01 -2.42996842e-01 -1.02551854e+00 1.13806057e+00
3.66775155e-01 5.00606120e-01 9.08670723e-01 4.33802724e-01
-8.22068453e-01 -8.87022316e-02 -1.15322936e+00 5.33580005e-01
2.62476146e-01 -8.62857759e-01 -5.27377903e-01 6.49268150e-01
4.18997467e-01 -4.69447821e-01 -1.20887983e+00 -1.01688191e-01
6.35666370e-01 -1.29646730e+00 1.01737499e+00 -5.03684878e-01
3.21388870e-01 -1.85495153e-01 1.55471712e-01 -1.18576264e+00
-7.66230226e-01 -6.90342724e-01 -3.09893101e-01 6.99550390e-01
2.94481605e-01 -5.70295572e-01 7.89889514e-01 7.84039199e-01
-4.01727527e-01 -4.28289443e-01 -5.86319506e-01 -7.71788120e-01
-2.66873483e-02 -7.32334256e-01 3.24064434e-01 1.02594686e+00
-6.50112778e-02 3.13864678e-01 6.57885522e-02 2.74371147e-01
7.76658237e-01 -1.73005685e-01 8.33560169e-01 -1.26353753e+00
-5.61544359e-01 -7.95290172e-01 -8.05355251e-01 -8.64050746e-01
-2.05820128e-01 -7.21472025e-01 -4.83442068e-01 -1.74626541e+00
1.73071951e-01 -2.72298902e-01 -3.43885332e-01 9.19504642e-01
8.23036358e-02 5.42030334e-01 4.13842387e-02 2.72479743e-01
-3.21804404e-01 1.49208605e-02 9.34179723e-01 3.06463968e-02
-1.28110856e-01 3.64362389e-01 -7.28258550e-01 8.78435969e-01
7.47590840e-01 -1.48865566e-01 -3.13940316e-01 -7.00932980e-01
-1.21806905e-01 -5.56382954e-01 8.86690736e-01 -1.23615026e+00
2.45196611e-01 -7.19005093e-02 6.75270379e-01 -2.71664292e-01
3.85646939e-01 -5.20855963e-01 -1.80900097e-01 2.71884888e-01
2.10104734e-02 2.42276073e-01 5.66351056e-01 -8.21378827e-03
-9.63257328e-02 1.06369220e-02 1.26655686e+00 -6.19158089e-01
-1.33073080e+00 6.22373283e-01 -3.37184429e-01 -1.21451408e-01
1.03614450e+00 -3.72369021e-01 -5.88084698e-01 -3.97387683e-01
-9.59150970e-01 1.84293270e-01 4.11222994e-01 5.91931462e-01
5.58713019e-01 -1.13502789e+00 -4.54727650e-01 4.12711293e-01
4.75217521e-01 3.60126048e-02 9.19632912e-02 7.05363393e-01
-8.08975577e-01 2.22279429e-01 -8.90682399e-01 -5.91203690e-01
-1.34154582e+00 3.62142742e-01 4.55841810e-01 4.99372035e-01
-9.52550411e-01 1.31379843e+00 -4.98369522e-02 -5.66710889e-01
4.43733156e-01 -4.12036598e-01 -2.27170475e-02 -1.45029962e-01
1.05247951e+00 5.57678342e-01 4.44639117e-01 -2.60904163e-01
-5.06133139e-01 2.45498970e-01 -2.25368127e-01 1.88435107e-01
1.31139660e+00 1.48977697e-01 -1.20942935e-01 7.18966663e-01
1.17170990e+00 -6.89725339e-01 -1.48477519e+00 -1.12925753e-01
-1.81931496e-01 -1.98979676e-01 3.55091631e-01 -7.31760323e-01
-9.24727321e-01 1.06483507e+00 6.51897609e-01 3.31737585e-02
8.22197378e-01 -7.41736516e-02 4.71190840e-01 5.49602747e-01
5.28060973e-01 -8.39248896e-01 7.71042630e-02 7.74368465e-01
8.18187773e-01 -1.19756663e+00 6.93396777e-02 8.43777284e-02
-6.57121301e-01 1.28312993e+00 8.19797516e-01 -1.36311471e-01
1.07439387e+00 5.60346544e-01 6.14098668e-01 -2.59555042e-01
-8.09940398e-01 -2.34380409e-01 1.64347038e-01 1.26399899e+00
7.95605063e-01 -1.80938020e-02 3.31357002e-01 1.51466593e-01
-4.18858111e-01 2.58199602e-01 5.19048810e-01 1.06136358e+00
-3.03192735e-01 -7.91220129e-01 -4.62069094e-01 4.30699646e-01
-4.80205476e-01 -1.53203890e-01 -4.53750379e-02 8.15036952e-01
6.85378090e-02 7.69424677e-01 3.36199790e-01 -4.11815912e-01
1.02974623e-01 -1.18557550e-01 3.85690570e-01 -7.71355331e-01
-7.55342603e-01 -2.34191105e-01 1.22248508e-01 -4.70717698e-01
5.06466702e-02 -3.95897388e-01 -1.01337922e+00 -7.33690500e-01
2.85198927e-01 -5.62257767e-02 1.20265090e+00 6.63252413e-01
3.75880152e-01 3.14622283e-01 3.33906531e-01 -1.37187386e+00
-3.62582862e-01 -1.20254803e+00 -6.98110700e-01 -1.38430834e-01
5.98302782e-01 -3.16217333e-01 -2.02058688e-01 1.99272528e-01]
|
[8.680054664611816, 2.7875308990478516]
|
70961c30-261c-494b-99a8-ce786a84ff14
|
lossy-compression-for-lossless-prediction
|
2106.10800
| null |
https://arxiv.org/abs/2106.10800v5
|
https://arxiv.org/pdf/2106.10800v5.pdf
|
Lossy Compression for Lossless Prediction
|
Most data is automatically collected and only ever "seen" by algorithms. Yet, data compressors preserve perceptual fidelity rather than just the information needed by algorithms performing downstream tasks. In this paper, we characterize the bit-rate required to ensure high performance on all predictive tasks that are invariant under a set of transformations, such as data augmentations. Based on our theory, we design unsupervised objectives for training neural compressors. Using these objectives, we train a generic image compressor that achieves substantial rate savings (more than $1000\times$ on ImageNet) compared to JPEG on 8 datasets, without decreasing downstream classification performance.
|
['Chris J. Maddison', 'Karen Ullrich', 'Benjamin Bloem-Reddy', 'Yann Dubois']
|
2021-06-21
| null |
http://proceedings.neurips.cc/paper/2021/hash/7535bbb91c8fde347ad861f293126633-Abstract.html
|
http://proceedings.neurips.cc/paper/2021/file/7535bbb91c8fde347ad861f293126633-Paper.pdf
|
neurips-2021-12
|
['feature-compression']
|
['computer-vision']
|
[ 7.19168544e-01 1.79059461e-01 -4.98157382e-01 -5.14337957e-01
-4.73274887e-01 -1.42062396e-01 4.24567521e-01 3.74090582e-01
-7.58328557e-01 3.77976686e-01 4.83739018e-01 -4.81067717e-01
1.16265118e-01 -6.79407179e-01 -1.07450473e+00 -4.28446919e-01
-4.95358169e-01 -2.44801827e-02 -1.65387079e-01 3.03413719e-02
3.28609943e-01 4.35853094e-01 -1.51097238e+00 4.37590688e-01
6.57856166e-01 1.19433641e+00 3.43522787e-01 9.17526007e-01
4.10076976e-01 1.11569703e+00 -4.46484655e-01 -5.07460952e-01
5.99753857e-01 -3.12809467e-01 -7.99025059e-01 6.21876121e-02
8.32633674e-01 -1.01223731e+00 -1.39547741e+00 1.12048268e+00
3.00082117e-01 3.96506369e-01 5.80285013e-01 -8.88080180e-01
-1.13671899e+00 8.29632461e-01 -8.28536153e-02 5.61630130e-01
-1.50454924e-01 4.61754739e-01 1.14344907e+00 -6.63397431e-01
3.85691494e-01 9.35761929e-01 7.09567368e-01 7.12996066e-01
-1.35157895e+00 -6.19511008e-01 -8.75721574e-02 4.42167848e-01
-1.22203064e+00 -1.00966299e+00 4.21130151e-01 -5.28913960e-02
1.38804758e+00 1.22552812e-01 7.16416001e-01 9.94635284e-01
3.44349533e-01 5.75323582e-01 6.26897097e-01 -1.97763264e-01
3.57276231e-01 -2.23500997e-01 -1.39708668e-01 7.06037223e-01
2.84320265e-01 3.12888265e-01 -9.94799674e-01 2.11893424e-01
5.93946934e-01 -1.80798784e-01 -5.30982792e-01 6.75983950e-02
-9.21944916e-01 6.88697100e-01 5.71245849e-01 -1.33554876e-01
-2.43986174e-01 7.60394871e-01 4.78030503e-01 6.73374832e-01
4.49431092e-01 6.09038174e-01 -3.99389774e-01 -3.00527602e-01
-1.11346507e+00 -3.50065306e-02 5.52466810e-01 1.31532526e+00
3.15976977e-01 4.99798536e-01 -7.33151510e-02 6.98897362e-01
1.15405478e-01 5.59139550e-01 8.40527415e-01 -1.37103367e+00
5.91797650e-01 3.62952910e-02 -3.54358554e-01 -8.95046949e-01
5.84304705e-03 -5.90986967e-01 -1.21543646e+00 6.25604670e-03
-6.27322644e-02 1.43864781e-01 -1.17297900e+00 2.04490924e+00
-4.49089468e-01 2.12784901e-01 1.66844040e-01 7.03220904e-01
4.41694170e-01 9.34366286e-01 2.91251659e-01 -1.55880660e-01
9.73851740e-01 -8.02216232e-01 -6.60422087e-01 -4.90525693e-01
6.71150327e-01 -4.73431766e-01 1.17950428e+00 5.44219613e-01
-1.49571717e+00 -6.98019326e-01 -1.65749025e+00 -5.30031443e-01
-3.17955986e-02 -2.27480128e-01 5.18380463e-01 5.62461197e-01
-1.46345413e+00 1.10244632e+00 -9.39445257e-01 8.15348923e-02
8.22467208e-01 5.08229792e-01 -2.99233973e-01 -2.38489240e-01
-8.61507952e-01 9.22454894e-01 5.01118362e-01 -4.09842700e-01
-1.28945494e+00 -9.55908239e-01 -7.91537702e-01 4.98439312e-01
-2.06306279e-01 -6.32276475e-01 1.60738039e+00 -8.59978199e-01
-1.21500611e+00 6.80309117e-01 5.63383028e-02 -1.29077268e+00
3.12038779e-01 -3.31892699e-01 -2.56046593e-01 5.28299153e-01
-3.39201301e-01 1.25782049e+00 1.05333459e+00 -1.14275086e+00
-7.91831017e-01 -1.85132399e-01 -1.68575004e-01 2.91117162e-01
-1.04926264e+00 -4.28371251e-01 -5.95605612e-01 -9.15049016e-01
-1.77594589e-03 -7.36709595e-01 -2.77487099e-01 3.38169456e-01
-1.62982062e-01 3.03085804e-01 1.04993546e+00 -8.83854151e-01
9.92353559e-01 -2.36549520e+00 8.29205438e-02 -4.59725894e-02
5.19251347e-01 3.41018617e-01 -4.68395323e-01 1.40227964e-02
2.28943136e-02 3.28968346e-01 -5.19171715e-01 -5.75276971e-01
-2.82081991e-01 3.06927085e-01 -5.45147121e-01 6.40851319e-01
1.18399888e-01 8.79849136e-01 -7.59782255e-01 -2.24634394e-01
1.32463396e-01 3.69373977e-01 -9.81224179e-01 3.43607754e-01
-2.82278746e-01 8.25615227e-02 1.25670493e-01 4.25950795e-01
6.08314216e-01 -4.32158232e-01 4.07941528e-02 -2.66697645e-01
3.23694676e-01 4.55119669e-01 -3.35184216e-01 1.89645243e+00
-5.76512516e-01 1.07926488e+00 -4.54653390e-02 -1.14670348e+00
6.65609896e-01 2.58029133e-01 5.42413771e-01 -8.75733137e-01
1.92264944e-01 6.28134683e-02 3.52623202e-02 -2.59999067e-01
6.79735482e-01 -4.10256833e-02 1.84517428e-01 3.54606569e-01
3.45983863e-01 -2.47254774e-01 -2.88584661e-02 2.59696662e-01
1.37616551e+00 -4.46547270e-01 8.67224783e-02 -1.71859860e-01
-8.39690864e-02 5.64796962e-02 4.14537668e-01 8.85474920e-01
-2.43865028e-01 3.32526535e-01 1.02871791e-01 -3.48241091e-01
-1.97224474e+00 -1.05341601e+00 -3.08455378e-01 1.06969857e+00
2.01688156e-01 -5.56948900e-01 -7.19576359e-01 -1.68231770e-01
1.51855916e-01 8.08336794e-01 -2.10268930e-01 -9.35645580e-01
-5.34305036e-01 -6.57203138e-01 1.02654886e+00 7.24523485e-01
8.44666719e-01 -8.61628234e-01 -7.78604031e-01 1.27328068e-01
-1.11470759e-01 -1.34754896e+00 -6.45275533e-01 6.42905235e-01
-1.38109732e+00 -5.75678229e-01 -3.12197477e-01 -8.34443748e-01
6.32765532e-01 3.91586900e-01 1.08359253e+00 1.51207343e-01
-4.30629738e-02 1.96867630e-01 -3.73803079e-01 -1.71505645e-01
-6.93494022e-01 1.17433108e-01 9.64330286e-02 -4.22232866e-01
2.47133046e-01 -9.04626548e-01 -7.10882723e-01 -2.88806915e-01
-1.02548993e+00 2.82335579e-01 9.07969117e-01 8.00112069e-01
6.82234645e-01 2.33810842e-01 3.28616828e-01 -6.81845605e-01
6.41263902e-01 -4.62546498e-01 -3.65898788e-01 -1.72129974e-01
-1.08214998e+00 3.74763429e-01 9.79272366e-01 -4.96558428e-01
-7.71245658e-01 6.99955523e-02 -1.60612881e-01 -8.75615001e-01
3.92611139e-02 4.65772301e-01 2.96722334e-02 -1.30007654e-01
8.70913506e-01 4.84175056e-01 8.31008330e-03 -4.07738149e-01
5.44318914e-01 6.45862997e-01 1.15490329e+00 -4.47463989e-01
7.08896697e-01 5.16817331e-01 7.22736865e-03 -8.42360675e-01
-7.32142389e-01 -2.18775552e-02 -4.63061243e-01 1.33681640e-01
7.43026972e-01 -1.08636653e+00 -5.04725218e-01 3.42069954e-01
-1.03436232e+00 -6.59401655e-01 -8.30966175e-01 4.65632051e-01
-8.06363881e-01 2.70628244e-01 -1.06423688e+00 -3.93328518e-01
-5.45625210e-01 -9.40312088e-01 4.93515790e-01 -1.33919328e-01
7.86693022e-02 -6.54441714e-01 -4.81411040e-01 1.51210845e-01
5.30988395e-01 -2.57985264e-01 1.08612144e+00 -2.94826150e-01
-7.53419936e-01 3.30168940e-02 -4.45118040e-01 9.17191148e-01
-3.76964286e-02 -3.56158227e-01 -1.14838672e+00 -5.67932487e-01
3.13237220e-01 -4.95689422e-01 1.33428383e+00 4.69573110e-01
2.08650708e+00 -9.90954220e-01 2.88279522e-02 1.39230835e+00
1.46451890e+00 3.07856172e-01 6.21527851e-01 2.15132181e-02
5.90673268e-01 2.24225685e-01 -1.02312220e-02 5.54096043e-01
-6.50491938e-02 1.51178792e-01 5.81893623e-01 1.13635786e-01
-4.49203253e-01 -6.49531782e-01 2.18823850e-01 9.79356170e-01
1.50971025e-01 -5.37121296e-01 -7.42427170e-01 5.00559628e-01
-1.32526386e+00 -8.52705359e-01 4.20902818e-01 1.98777831e+00
1.09419346e+00 4.38984752e-01 -3.99973541e-01 3.02828521e-01
5.20382464e-01 3.46692771e-01 -9.39750612e-01 -4.93474871e-01
-8.89205486e-02 5.43682396e-01 1.28939486e+00 4.20675576e-01
-1.12601745e+00 8.69288802e-01 7.68524265e+00 6.20325983e-01
-8.92433643e-01 1.37849182e-01 1.11591637e+00 -3.98525417e-01
-3.25480700e-01 -1.25263438e-01 -3.74541879e-01 4.20332462e-01
1.55678070e+00 -3.51136029e-01 8.50689769e-01 1.01987934e+00
2.15113774e-01 4.08515215e-01 -1.54049122e+00 1.10413325e+00
9.61116403e-02 -1.66538906e+00 5.30295372e-01 2.12833300e-01
8.45604837e-01 1.90544456e-01 5.47930896e-01 6.95764869e-02
4.85247195e-01 -1.36405385e+00 9.09821749e-01 9.47896466e-02
1.41993594e+00 -7.76859701e-01 4.16704208e-01 2.60915697e-01
-7.92336166e-01 -5.40936053e-01 -6.48788869e-01 2.35293936e-02
3.55340093e-02 5.26511669e-01 -6.77334487e-01 -3.03994924e-01
6.76101923e-01 9.05882776e-01 -4.09666777e-01 1.02129006e+00
-7.59255737e-02 6.20485783e-01 -1.02592133e-01 4.24766123e-01
1.48572832e-01 2.63663709e-01 2.73612291e-01 1.40523124e+00
5.36555290e-01 3.05031002e-01 -1.58005014e-01 5.11747777e-01
-8.51991355e-01 -3.51867676e-01 -8.91664445e-01 -1.34761602e-01
7.94558883e-01 5.34348309e-01 -1.84019029e-01 -5.08278370e-01
-2.75613517e-01 1.22967362e+00 2.40521178e-01 2.50246793e-01
-5.77585399e-01 -1.59461483e-01 9.09348845e-01 2.24551499e-01
2.35418335e-01 -5.42049944e-01 -6.55866027e-01 -9.90703166e-01
-1.11527607e-01 -1.04413879e+00 2.65861988e-01 -6.82400286e-01
-1.15007889e+00 3.72286230e-01 -9.30095315e-02 -1.22584796e+00
-3.38996165e-02 -6.59931660e-01 -1.60570532e-01 4.11220610e-01
-1.62108934e+00 -7.42756665e-01 -3.30541372e-01 6.18126214e-01
7.42829442e-01 -2.87743002e-01 7.52730668e-01 4.63899672e-01
-1.73349172e-01 9.52392340e-01 1.85918704e-01 7.53530562e-02
3.40078264e-01 -1.12283492e+00 6.89337313e-01 1.20612729e+00
2.39377335e-01 4.54693466e-01 6.91098154e-01 -4.71587270e-01
-1.52734983e+00 -1.27633572e+00 5.58694899e-01 1.82525799e-01
4.62051421e-01 -1.30290881e-01 -8.85412335e-01 8.78082037e-01
3.43645006e-01 1.95141539e-01 6.06033683e-01 -2.66278595e-01
-7.26957500e-01 -3.21371317e-01 -1.14471900e+00 5.09722054e-01
1.43919957e+00 -7.98586726e-01 -2.66997188e-01 3.28598380e-01
1.23743594e+00 -5.61256230e-01 -1.03715026e+00 4.32818025e-01
3.18468243e-01 -6.49117768e-01 1.08762860e+00 -8.55577648e-01
9.80955958e-01 3.16861600e-01 -6.01145029e-01 -1.31533301e+00
-3.73566955e-01 -6.14117384e-01 -7.30892599e-01 5.16996682e-01
2.82996148e-01 -2.48590112e-01 1.02258790e+00 5.00591755e-01
-6.27646685e-01 -4.53214139e-01 -1.07181418e+00 -7.12209523e-01
2.54592318e-02 -5.73432684e-01 3.35133165e-01 7.98817992e-01
-3.36373560e-02 9.15383101e-02 -5.26493490e-01 1.33596912e-01
7.02280700e-01 -3.87872517e-01 2.63468891e-01 -6.96551383e-01
-5.00403106e-01 -5.84211588e-01 -5.37566662e-01 -1.61147559e+00
-1.22925648e-02 -1.01196957e+00 2.17377022e-01 -1.07037568e+00
2.87471771e-01 -6.89440131e-01 -3.76913816e-01 4.81837511e-01
3.04213017e-01 2.37991780e-01 2.36333728e-01 4.70477670e-01
-2.23787189e-01 8.25926781e-01 1.26544905e+00 -4.16358769e-01
4.22893167e-02 -3.34627837e-01 -9.73720074e-01 5.69901824e-01
9.39205468e-01 -4.85185087e-01 -8.47217262e-01 -1.16424716e+00
-1.14837967e-01 8.31631124e-02 2.29804084e-01 -1.45203280e+00
3.14864933e-01 -2.01179832e-01 5.49492717e-01 -2.57541567e-01
5.27307212e-01 -7.14397132e-01 -3.02909493e-01 8.39043140e-01
-1.05855453e+00 3.08156639e-01 1.13991715e-01 7.40910053e-01
-1.53281912e-01 -3.11828494e-01 1.15626240e+00 -4.40832712e-02
-7.36125052e-01 6.18958712e-01 -3.67298156e-01 3.12355496e-02
8.86279762e-01 -4.86245984e-03 -5.85407436e-01 -6.73025846e-01
-5.58036268e-01 -2.09230512e-01 4.58424628e-01 1.40064776e-01
1.07789648e+00 -1.27218938e+00 -7.48381674e-01 4.10918206e-01
-8.05791393e-02 7.76622966e-02 -9.10910070e-02 1.22670017e-01
-8.94817054e-01 4.89712834e-01 -4.93708044e-01 -3.05027366e-01
-9.76296365e-01 7.48533726e-01 3.32379162e-01 1.11833125e-01
-1.03722680e+00 9.15879250e-01 -1.74341843e-01 2.72527844e-01
6.06581569e-01 -5.21227658e-01 2.68222272e-01 -4.72250611e-01
5.04849136e-01 7.01382682e-02 9.77988392e-02 -1.61991000e-01
2.15848789e-01 -1.22390069e-01 -1.14749059e-01 -2.05181837e-01
1.36855614e+00 -8.43957067e-02 3.48055184e-01 -1.50308218e-02
1.51858354e+00 -4.42194700e-01 -1.70859075e+00 -3.05524439e-01
-2.18148768e-01 -7.13157833e-01 5.39134324e-01 -3.67661238e-01
-1.38409328e+00 8.96677375e-01 7.51612008e-01 1.53238505e-01
1.50808179e+00 -1.95824981e-01 1.16826642e+00 7.41332412e-01
1.17814153e-01 -1.18007600e+00 2.71705359e-01 3.12594235e-01
6.27799869e-01 -1.01108539e+00 3.61079164e-02 -1.96602046e-01
-4.38953996e-01 8.13165188e-01 6.52602196e-01 -3.39573205e-01
7.46436059e-01 6.50242150e-01 -5.22967100e-01 9.18054432e-02
-1.13051999e+00 1.93914115e-01 3.16897221e-02 7.26916015e-01
1.44978225e-01 -4.22955863e-02 1.57178473e-02 -6.37036264e-02
-5.76189578e-01 -7.57146347e-03 5.29075027e-01 9.15016532e-01
-7.74676919e-01 -7.44138122e-01 1.52416542e-01 8.44567060e-01
-4.10671175e-01 -4.43622112e-01 1.14080906e-01 4.30977523e-01
8.64321813e-02 1.00972199e+00 5.85419178e-01 -8.97357047e-01
-5.85432425e-02 -2.58752048e-01 4.73663688e-01 -3.61464858e-01
-8.72165486e-02 -4.01839733e-01 1.23327136e-01 -5.12947440e-01
-2.11762652e-01 -3.80930811e-01 -1.09736776e+00 -1.10789812e+00
1.02005184e-01 -2.63303548e-01 7.07914054e-01 6.70302033e-01
3.02011937e-01 5.64560533e-01 6.21640086e-01 -6.32651687e-01
-9.46045876e-01 -8.47900808e-01 -3.46483648e-01 5.80215096e-01
4.04069692e-01 1.82625949e-02 -3.67731065e-01 7.15749681e-01]
|
[11.400357246398926, -1.5583937168121338]
|
f7c07ed8-fa81-44ea-84b4-c290135121af
|
graph-based-long-term-and-short-term-interest
|
2306.10028
| null |
https://arxiv.org/abs/2306.10028v1
|
https://arxiv.org/pdf/2306.10028v1.pdf
|
Graph Based Long-Term And Short-Term Interest Model for Click-Through Rate Prediction
|
Click-through rate (CTR) prediction aims to predict the probability that the user will click an item, which has been one of the key tasks in online recommender and advertising systems. In such systems, rich user behavior (viz. long- and short-term) has been proved to be of great value in capturing user interests. Both industry and academy have paid much attention to this topic and propose different approaches to modeling with long-term and short-term user behavior data. But there are still some unresolved issues. More specially, (1) rule and truncation based methods to extract information from long-term behavior are easy to cause information loss, and (2) single feedback behavior regardless of scenario to extract information from short-term behavior lead to information confusion and noise. To fill this gap, we propose a Graph based Long-term and Short-term interest Model, termed GLSM. It consists of a multi-interest graph structure for capturing long-term user behavior, a multi-scenario heterogeneous sequence model for modeling short-term information, then an adaptive fusion mechanism to fused information from long-term and short-term behaviors. Comprehensive experiments on real-world datasets, GLSM achieved SOTA score on offline metrics. At the same time, the GLSM algorithm has been deployed in our industrial application, bringing 4.9% CTR and 4.3% GMV lift, which is significant to the business.
|
['Dong Wang', 'Xingxing Wang', 'Bo Zhang', 'Pengye Zhang', 'Guangliang Yu', 'Huinan Sun']
|
2023-06-05
| null | null | null | null |
['click-through-rate-prediction']
|
['miscellaneous']
|
[-6.21936172e-02 -5.40252626e-01 -5.46184897e-01 -4.21891779e-01
-3.43853235e-01 -3.42237592e-01 3.76033902e-01 1.39307871e-01
-6.41051456e-02 4.98512477e-01 2.47008398e-01 -4.41217959e-01
-4.17839080e-01 -8.89242291e-01 -4.15390968e-01 -3.15877438e-01
-2.62523085e-01 1.04137614e-01 4.23867643e-01 -6.91141307e-01
4.04358774e-01 1.30882384e-02 -1.71376491e+00 3.84001344e-01
9.72220778e-01 1.35934019e+00 3.82456750e-01 4.43442076e-01
-4.30416703e-01 5.89848280e-01 -3.44867885e-01 -6.05980694e-01
3.04671586e-01 -2.77198225e-01 -3.43786448e-01 1.05916271e-02
-1.59775674e-01 -2.02757061e-01 -3.06558192e-01 9.96960938e-01
4.41689819e-01 1.47606865e-01 2.54082233e-01 -1.24238360e+00
-6.33097887e-01 6.60205722e-01 -7.38201857e-01 2.98058897e-01
5.37680030e-01 -3.39581631e-02 1.11103797e+00 -6.36428595e-01
4.28029805e-01 1.06873369e+00 3.53656113e-01 6.26291111e-02
-8.61190140e-01 -6.83422148e-01 6.11839890e-01 4.02196884e-01
-1.18142092e+00 6.11336529e-02 7.17773139e-01 -3.28313857e-01
4.22281474e-01 5.74654102e-01 8.26649785e-01 8.92865062e-01
5.31796455e-01 1.00473475e+00 1.00795352e+00 -1.13609750e-02
-3.54034407e-03 5.07400572e-01 7.04155028e-01 3.28569531e-01
1.56679854e-01 7.93567598e-02 -1.46882981e-01 -2.94762045e-01
3.98636997e-01 5.26300490e-01 -1.07349299e-01 8.20117667e-02
-6.51074409e-01 9.39224899e-01 4.51277822e-01 4.57662851e-01
-3.71584952e-01 -5.21324158e-01 1.32850811e-01 5.81966281e-01
5.24483860e-01 1.54141579e-02 -4.54482138e-01 -1.54467583e-01
-6.96731329e-01 2.57649899e-01 9.89489257e-01 1.08147967e+00
6.20899856e-01 -2.70566251e-02 -3.04383337e-01 9.59889770e-01
4.80130494e-01 3.89210790e-01 7.49206841e-01 -2.73041815e-01
5.02806723e-01 8.35735619e-01 1.69080421e-01 -1.41278887e+00
-3.76460820e-01 -1.15496731e+00 -7.34881520e-01 -3.92780393e-01
-2.69391900e-03 -3.41767743e-02 -6.36110187e-01 1.38409209e+00
2.13250831e-01 1.57998413e-01 -3.16603124e-01 9.86051023e-01
7.03296602e-01 7.63496280e-01 -8.82029235e-02 -6.75518215e-01
1.32409513e+00 -8.10796142e-01 -9.07801270e-01 -1.73399657e-01
3.95949692e-01 -9.92670894e-01 9.96521115e-01 5.61787367e-01
-9.43310082e-01 -7.54576743e-01 -9.69458520e-01 4.34299290e-01
-5.03329396e-01 1.72258392e-01 8.38030636e-01 5.79062939e-01
-4.66941088e-01 3.41010511e-01 -1.15127511e-01 -5.10916486e-02
-3.99471447e-02 3.43534112e-01 2.83903569e-01 -1.72827005e-01
-1.53966820e+00 4.68835562e-01 1.40176073e-01 1.09959066e-01
-4.73583251e-01 -6.11854553e-01 -2.77100027e-01 1.86650306e-01
7.51140475e-01 -2.88618714e-01 1.12067676e+00 -6.38934016e-01
-1.18570161e+00 -3.06443144e-02 -2.40273967e-01 -3.69913876e-01
3.08184236e-01 -4.56450954e-02 -1.46630478e+00 -4.60085452e-01
-8.70506689e-02 -4.13599730e-01 6.88603461e-01 -9.44923222e-01
-9.92992997e-01 -7.40897477e-01 4.96414155e-02 2.39887878e-01
-4.15822953e-01 -9.50217098e-02 -7.42998242e-01 -7.54914582e-01
1.40558705e-01 -9.04687524e-01 -3.29899609e-01 -6.36269331e-01
-1.58416778e-01 -2.91414618e-01 7.96285987e-01 -7.03456104e-01
2.04601312e+00 -1.89372754e+00 -4.28130984e-01 3.68572950e-01
2.54916161e-01 6.00542486e-01 4.30290513e-02 5.83933055e-01
3.31615061e-01 8.93498734e-02 4.69646096e-01 4.01595294e-01
-1.37445837e-01 -3.86699364e-02 -3.02761465e-01 -1.92322470e-02
-4.97403026e-01 8.22226584e-01 -7.60283053e-01 -2.67469704e-01
2.69232452e-01 2.47254640e-01 -4.97773558e-01 1.49459943e-01
-3.32722098e-01 1.91240698e-01 -7.95046747e-01 6.18391037e-01
8.67732942e-01 -6.07470632e-01 -1.82949076e-03 -3.58188450e-01
-1.48838326e-01 1.16308434e-02 -1.31191540e+00 1.07892275e+00
-5.69209039e-01 -1.34796441e-01 9.83068049e-02 -6.89958930e-01
9.62956548e-01 -4.88661118e-02 6.56934142e-01 -1.12169683e+00
3.98812205e-01 1.23094358e-01 4.50867554e-03 -5.86829007e-01
6.29259884e-01 2.51205593e-01 -6.76631415e-03 3.82670671e-01
-1.02840170e-01 6.20753765e-01 3.74857754e-01 4.59404677e-01
7.62552381e-01 -1.91530287e-01 2.14539543e-01 8.91336706e-03
7.69064426e-01 -5.87981082e-02 4.80143309e-01 5.49576283e-01
-4.77044582e-02 2.46228203e-01 -6.66114911e-02 -6.82830438e-02
-5.37150621e-01 -6.37761176e-01 1.28758296e-01 1.33256209e+00
5.94625592e-01 -6.76245749e-01 -1.07609384e-01 -7.01055050e-01
-5.65293618e-03 7.37863421e-01 -2.14042664e-01 -2.44570643e-01
-3.16397935e-01 -6.62935257e-01 -1.63457647e-01 1.04409918e-01
4.97607261e-01 -5.67020595e-01 3.75353366e-01 5.62373221e-01
-1.94258437e-01 -9.21636820e-01 -8.18196356e-01 -3.81176084e-01
-7.41165698e-01 -9.46202934e-01 -7.39061475e-01 -4.52956021e-01
1.98907241e-01 8.57087553e-01 9.18913603e-01 -2.94336304e-02
-1.50091156e-01 1.00051358e-01 -6.19588852e-01 -2.92372584e-01
3.14621590e-02 -1.03288621e-01 -1.82295829e-01 5.17134368e-01
6.42850935e-01 -5.14801800e-01 -1.02788651e+00 1.02320850e+00
-7.57984698e-01 -2.94590890e-01 7.80318558e-01 5.92582524e-01
4.81512636e-01 1.15654416e-01 7.41948485e-01 -9.88446653e-01
9.77043629e-01 -1.07323754e+00 -6.86305463e-01 3.69006217e-01
-1.22103179e+00 -2.62916178e-01 8.71211708e-01 -4.51774448e-01
-1.16422105e+00 -2.90922523e-01 -3.81348938e-01 1.78451911e-02
2.00635597e-01 7.60686040e-01 -8.06558225e-03 -1.02793343e-01
4.38333243e-01 5.56511164e-01 -8.37107655e-03 -7.54985809e-01
4.49737668e-01 1.14619315e+00 -3.27639319e-02 -7.07640871e-02
3.18094909e-01 1.75340533e-01 -8.15427601e-02 -7.12651908e-01
-9.78563964e-01 -1.07444358e+00 5.61656170e-02 -4.51596349e-01
4.40972179e-01 -9.22336519e-01 -1.03198874e+00 1.61960930e-01
-5.08153319e-01 6.06670380e-01 1.70129046e-01 7.32880056e-01
-1.34146720e-01 4.54630494e-01 -5.52729011e-01 -9.77597117e-01
-5.16925395e-01 -1.05557144e+00 5.15089512e-01 4.29609656e-01
2.44051427e-01 -9.17240620e-01 -2.44680658e-01 4.85574722e-01
6.91567361e-01 -3.00605714e-01 5.07257760e-01 -9.31051910e-01
-5.63320458e-01 -5.29966474e-01 -2.88461596e-01 3.41021121e-01
1.26059651e-01 -4.15632248e-01 -3.68208230e-01 -3.34171563e-01
2.09871978e-01 7.15083554e-02 5.05404830e-01 2.82390714e-01
1.13750863e+00 -4.33004379e-01 -4.67582554e-01 2.27727070e-01
1.43091512e+00 5.75616419e-01 4.27214354e-01 -1.55573413e-01
5.12225807e-01 3.54669899e-01 1.16446173e+00 6.69844985e-01
3.84814054e-01 8.30170095e-01 3.50299269e-01 2.62400180e-01
-2.01007202e-02 -5.07491231e-01 2.99401939e-01 1.17313778e+00
2.49724146e-02 -4.53511000e-01 -1.28710032e-01 1.56620115e-01
-1.99985552e+00 -1.02466762e+00 -5.86901605e-01 2.35158420e+00
3.65872532e-01 2.90235400e-01 3.56569290e-01 4.13365513e-02
8.61943722e-01 1.42092496e-01 -6.22850418e-01 -1.37253255e-01
6.80676624e-02 -2.61083543e-01 6.13293350e-01 3.37383151e-01
-6.26594067e-01 5.33795238e-01 5.50807333e+00 1.27941144e+00
-1.04848075e+00 5.01036532e-02 4.93453354e-01 2.11059302e-02
-4.24465656e-01 4.87454981e-03 -1.22614241e+00 9.55671847e-01
9.48818266e-01 -5.72435558e-01 4.46157545e-01 8.52536380e-01
3.75874072e-01 1.36578619e-01 -4.06144887e-01 9.99015450e-01
1.48217931e-01 -9.82760966e-01 1.32279828e-01 3.53155106e-01
5.55908978e-01 -1.55841991e-01 4.10739109e-02 6.71787262e-01
3.79749000e-01 -4.53044504e-01 1.85751647e-01 6.73183978e-01
3.89863074e-01 -6.22934699e-01 7.60986090e-01 6.51072681e-01
-1.50506067e+00 -3.85771751e-01 -3.83451730e-01 -1.00454077e-01
4.01372999e-01 1.05963159e+00 -4.96757835e-01 1.03986192e+00
3.83519351e-01 7.87519932e-01 -5.11072814e-01 1.29977679e+00
2.69619823e-01 7.56404519e-01 -3.19880247e-01 -5.66744804e-01
1.57877222e-01 -6.22442603e-01 4.11107987e-01 7.85230517e-01
5.61605215e-01 1.87124953e-01 4.79146510e-01 4.04625118e-01
1.52869880e-01 5.60760260e-01 -3.03639859e-01 -1.11829653e-01
2.07781196e-01 1.38766932e+00 -2.90698349e-01 -3.46164823e-01
-6.28140748e-01 5.50420046e-01 -1.33709654e-01 3.57093483e-01
-9.12264407e-01 -4.01417434e-01 2.37069860e-01 4.84618276e-01
3.14476460e-01 9.87756625e-03 -5.04653081e-02 -1.23056054e+00
1.28401890e-01 -6.58962786e-01 7.59245872e-01 -5.45291424e-01
-1.37662613e+00 5.29410541e-01 -2.51580298e-01 -1.70958257e+00
-1.59745097e-01 -2.49211207e-01 -3.58735621e-01 8.40069592e-01
-1.40096164e+00 -9.86863792e-01 -4.42800701e-01 8.59006107e-01
6.85549974e-01 -2.17171565e-01 3.27746570e-01 7.58709371e-01
-3.46018285e-01 6.80167913e-01 2.27104887e-01 -3.78340036e-01
3.99064481e-01 -7.00710058e-01 4.61678244e-02 3.60732704e-01
9.29303840e-02 9.42953229e-01 7.29564786e-01 -7.57214248e-01
-1.70809984e+00 -1.17576504e+00 7.94537008e-01 8.36271271e-02
6.98827565e-01 -2.76367873e-01 -7.22667217e-01 4.77635950e-01
-2.04823688e-01 -1.19295344e-01 7.08153784e-01 3.84903938e-01
-9.41153243e-02 -4.46401715e-01 -1.06317115e+00 6.09472811e-01
1.06714296e+00 -1.04143824e-02 -2.83685327e-01 5.57744861e-01
8.19577813e-01 8.25391635e-02 -1.02650118e+00 2.72439480e-01
8.02369654e-01 -1.11594892e+00 8.54886293e-01 -4.63017493e-01
-2.85088327e-02 -2.38077655e-01 -2.29643971e-01 -1.09115493e+00
-6.87975347e-01 -5.28480291e-01 -4.69979942e-01 1.17835474e+00
4.48455811e-01 -6.48893476e-01 8.72976005e-01 4.32829499e-01
-5.93354413e-03 -8.49519908e-01 -2.94127733e-01 -8.95658135e-01
-5.11985064e-01 -4.32803929e-01 6.18943214e-01 5.80087960e-01
1.11064166e-01 7.74844587e-01 -9.20876801e-01 -9.46279243e-02
5.09243011e-01 6.14054680e-01 6.93236232e-01 -1.34954321e+00
-4.15079981e-01 -3.11285764e-01 -2.82201588e-01 -1.64775205e+00
-5.40756524e-01 -7.28690624e-01 -5.17397165e-01 -1.48897612e+00
1.83664292e-01 -4.27747041e-01 -5.68682611e-01 -8.39233771e-02
-1.09862618e-01 -1.47551652e-02 4.12105359e-02 2.81655312e-01
-8.06490302e-01 4.50389296e-01 1.34412456e+00 -1.20939128e-02
-3.71521026e-01 9.25261199e-01 -8.81517351e-01 3.47310364e-01
7.42917657e-01 -1.25429317e-01 -6.98334813e-01 1.75483227e-01
3.62125009e-01 5.52163541e-01 -2.24536419e-01 -8.27027857e-01
3.90689969e-01 -3.79725039e-01 9.03637856e-02 -8.93655419e-01
5.01061738e-01 -1.02397394e+00 2.77805448e-01 4.36031997e-01
-3.39750290e-01 -1.48210481e-01 -3.22900265e-01 1.07678783e+00
-3.45174938e-01 -9.29182395e-02 4.07202512e-01 -1.65902540e-01
-8.49418044e-01 6.86502934e-01 -7.53613412e-02 -1.70539513e-01
1.07795584e+00 -5.43932430e-02 -3.26796889e-01 -7.88476169e-01
-8.91132593e-01 5.45737863e-01 -1.77962139e-01 8.30295146e-01
4.76942420e-01 -1.40801239e+00 -5.46300888e-01 2.71741092e-01
3.54891151e-01 -9.75572288e-01 6.44646287e-01 7.82301009e-01
1.39910340e-01 7.21804738e-01 1.79559514e-01 -3.78462374e-01
-1.39062655e+00 8.37005734e-01 -3.19697261e-01 -4.45414066e-01
-9.71112922e-02 4.82261270e-01 -1.53791577e-01 -1.19203925e-01
1.30946845e-01 -1.74756926e-02 -7.60366797e-01 1.57346949e-01
6.85096860e-01 6.11383200e-01 2.87061602e-01 -6.85038328e-01
-2.46773120e-02 5.31949043e-01 -4.99755710e-01 1.62613213e-01
9.13506806e-01 -6.55887306e-01 2.41060570e-01 3.63846064e-01
1.41167557e+00 1.98566332e-01 -5.10612845e-01 -5.42566121e-01
-2.69221604e-01 -7.54308105e-01 2.62208611e-01 -1.09692895e+00
-1.21512318e+00 5.91412425e-01 6.52099788e-01 8.23709428e-01
1.08722889e+00 -2.74727553e-01 1.32913446e+00 1.24954320e-01
7.65779793e-01 -1.12695050e+00 -1.61210552e-01 2.49893770e-01
7.26626098e-01 -1.24319696e+00 1.93149626e-01 -6.93469703e-01
-7.98218191e-01 7.32840359e-01 4.99209434e-01 -3.27660181e-02
1.26228857e+00 -2.19140783e-01 -2.77730078e-01 -8.69998932e-02
-9.52876389e-01 -4.52653199e-01 5.87403178e-01 1.83024436e-01
4.07770306e-01 1.53397292e-01 -1.07874322e+00 9.87383306e-01
-8.86109471e-02 3.38927329e-01 2.24151075e-01 8.74057710e-01
-6.13644183e-01 -1.18809831e+00 2.99122483e-02 1.15603173e+00
-6.70513749e-01 -5.62325306e-02 2.32344493e-02 3.29371780e-01
-9.75624993e-02 1.50013196e+00 -4.77529407e-01 -1.02923667e+00
4.37634468e-01 -3.14345688e-01 -5.08780181e-02 -2.83516139e-01
-6.55344963e-01 4.93470639e-01 1.73663378e-01 -8.37571144e-01
2.60049887e-02 -3.36013734e-01 -8.52832496e-01 -3.01215857e-01
-8.34586024e-01 5.93583405e-01 8.88080418e-01 6.65604830e-01
7.90741503e-01 5.13062835e-01 1.14921737e+00 -1.21540539e-01
-9.08315182e-01 -1.02788913e+00 -9.61254835e-01 6.65251255e-01
-8.50554705e-02 -5.96398115e-01 -4.85336632e-01 -4.60715771e-01]
|
[10.111244201660156, 5.602616786956787]
|
952f230f-e92b-420c-a364-a21af08ea795
|
toward-qualitative-evaluation-of-embeddings
| null | null |
https://aclanthology.org/2020.lrec-1.610
|
https://aclanthology.org/2020.lrec-1.610.pdf
|
Toward Qualitative Evaluation of Embeddings for Arabic Sentiment Analysis
|
In this paper, we propose several protocols to evaluate specific embeddings for Arabic sentiment analysis (SA) task. In fact, Arabic language is characterized by its agglutination and morphological richness contributing to great sparsity that could affect embedding quality. This work presents a study that compares embeddings based on words and lemmas in SA frame. We propose first to study the evolution of embedding models trained with different types of corpora (polar and non polar) and explore the variation between embeddings by observing the sentiment stability of neighbors in embedding spaces. Then, we evaluate embeddings with a neural architecture based on convolutional neural network (CNN). We make available our pre-trained embeddings to Arabic NLP research community with free to use. We provide also for free resources used to evaluate our embeddings. Experiments are done on the Large Arabic-Book Reviews (LABR) corpus in binary (positive/negative) classification frame. Our best result reaches 91.9{\%}, that is higher than the best previous published one (91.5{\%}).
|
['lamia hadrich belguith', 'Yannick Est{\\`e}ve', 'Amira Barhoumi', 'Chafik Aloulou', 'Nathalie Camelin']
|
2020-05-01
| null | null | null |
lrec-2020-5
|
['arabic-sentiment-analysis']
|
['natural-language-processing']
|
[-3.88978422e-01 -6.53991401e-02 -8.91862810e-02 -4.12250876e-01
-2.82492459e-01 -8.53870273e-01 6.36970460e-01 6.88776433e-01
-7.13306308e-01 5.22764325e-01 3.26475173e-01 -3.55862886e-01
5.72944619e-02 -9.12160754e-01 -4.06778306e-01 -6.62885368e-01
-4.19561267e-01 4.30052549e-01 2.17033029e-02 -1.04616380e+00
4.74668860e-01 5.96430063e-01 -1.38395262e+00 3.41928184e-01
7.76698768e-01 9.29276109e-01 -2.38789424e-01 6.43706620e-01
-1.34149790e-01 4.04538304e-01 -7.25729883e-01 -8.63690317e-01
5.98189086e-02 -7.84212723e-02 -7.51237988e-01 -2.76044816e-01
4.29135039e-02 1.48791134e-01 1.89161316e-01 1.01782990e+00
4.52056885e-01 1.43053174e-01 1.07985258e+00 -1.05162477e+00
-1.37648773e+00 7.17606187e-01 -3.02021503e-01 2.87795484e-01
3.03249776e-01 -1.93366557e-01 1.34273326e+00 -1.16488254e+00
6.78459048e-01 1.01397443e+00 6.81601584e-01 3.28228801e-01
-5.80151021e-01 -1.43438682e-01 1.68801956e-02 2.84078449e-01
-1.12339699e+00 -1.90875590e-01 8.80528152e-01 -2.31638610e-01
1.09537435e+00 1.66497633e-01 7.32806563e-01 1.07729042e+00
2.88073987e-01 5.55756032e-01 1.20872319e+00 -8.53563190e-01
1.29191950e-01 5.78895211e-01 5.53627908e-01 6.68664932e-01
2.87626475e-01 -5.01442969e-01 -3.51531059e-01 5.45333885e-03
4.66400310e-02 -3.98623407e-01 -1.73293814e-01 -9.56956446e-02
-1.08491516e+00 1.18335032e+00 5.12406111e-01 6.93917274e-01
-2.72816390e-01 -2.27394924e-01 5.94985366e-01 4.90715533e-01
4.89180416e-01 6.13867879e-01 -6.79302514e-01 -3.03279281e-01
-4.82723087e-01 9.75894034e-02 1.00038886e+00 4.68344361e-01
4.73158687e-01 1.89193249e-01 2.34594613e-01 1.12518358e+00
6.26116991e-01 4.49639440e-01 9.21820283e-01 -2.41768584e-01
2.18531162e-01 6.44690871e-01 6.32416010e-02 -1.45879984e+00
-4.59090084e-01 -2.87796646e-01 -5.21937370e-01 8.81949291e-02
5.39175749e-01 -4.06703591e-01 -6.38182998e-01 1.21644258e+00
4.15309034e-02 -5.74742675e-01 5.46869516e-01 8.75657797e-01
9.52767491e-01 9.79004204e-01 -2.53290176e-01 1.29426330e-01
1.81535399e+00 -1.06821167e+00 -9.13945496e-01 -1.18611209e-01
8.48025858e-01 -1.04868925e+00 1.29505491e+00 5.84787786e-01
-8.79826903e-01 -2.90072232e-01 -1.54991996e+00 -1.44703060e-01
-1.33235574e+00 3.92689973e-01 4.13012534e-01 1.13114429e+00
-1.12295175e+00 5.46784282e-01 -7.00648665e-01 -5.70906520e-01
1.90956056e-01 4.84193355e-01 -6.52305484e-01 1.66734040e-01
-1.44292140e+00 1.45411599e+00 3.73458862e-01 3.60390663e-01
-3.54567081e-01 -3.27885300e-02 -1.09072685e+00 -1.84984058e-01
-2.96848089e-01 3.06880444e-01 6.26482666e-01 -1.16053486e+00
-1.53687024e+00 9.27175820e-01 4.76966910e-02 -3.90987694e-01
-8.55835006e-02 -3.61912191e-01 -8.41984391e-01 2.08558664e-02
-2.65293121e-01 6.18948102e-01 4.94533330e-01 -1.20223856e+00
-1.52677268e-01 -4.20161396e-01 3.78076345e-01 8.97244141e-02
-1.09222806e+00 1.67062908e-01 -4.14847210e-02 -6.66149557e-01
-2.44498681e-02 -9.90886688e-01 7.94004500e-02 -4.34916079e-01
-2.69086491e-02 -4.18729484e-01 6.63871169e-01 -7.32537329e-01
1.30452931e+00 -2.15931296e+00 1.08644046e-01 1.71017140e-01
-1.99677974e-01 5.17892897e-01 -3.41205120e-01 7.29042351e-01
-2.21441984e-01 2.58477420e-01 -3.07666481e-01 -2.50169128e-01
2.47983634e-01 2.69796461e-01 -1.39462546e-01 6.98271394e-01
6.74040854e-01 6.73580468e-01 -8.26254606e-01 -3.40485722e-01
2.49163080e-02 7.69179940e-01 -4.04460281e-01 -1.85517415e-01
6.94240630e-02 -2.46963114e-01 -1.08196221e-01 9.04239416e-01
7.29627609e-01 1.37805313e-01 3.07681441e-01 -4.08187568e-01
-2.72360981e-01 2.78849959e-01 -1.11298811e+00 1.33683968e+00
-5.08283675e-01 9.40153182e-01 -2.06465214e-01 -9.97214556e-01
1.40670466e+00 1.42850637e-01 1.32607073e-01 -4.48112458e-01
5.28960288e-01 5.05062938e-01 2.97775269e-01 -3.20474565e-01
1.09604728e+00 1.28992982e-02 -6.28552539e-03 4.93363380e-01
4.27416801e-01 6.85903663e-03 4.99970794e-01 1.56078935e-02
6.20212257e-01 -1.45753354e-01 1.62388474e-01 -6.79449737e-01
8.90309095e-01 5.98454028e-02 -9.36961025e-02 5.21016568e-02
-4.87467825e-01 5.88816285e-01 1.04606557e+00 -6.32718801e-01
-9.80173767e-01 -7.95229971e-01 -5.53006887e-01 1.14947140e+00
-9.70681459e-02 -5.28159738e-01 -5.82280159e-01 -7.68697023e-01
-3.73047531e-01 5.55035174e-01 -1.09410107e+00 -2.66308188e-02
-6.16215050e-01 -1.33706808e+00 5.28373659e-01 4.44010675e-01
1.86822396e-02 -1.31960380e+00 -3.64364624e-01 -2.74933446e-02
1.95529014e-01 -7.49992728e-01 4.99048159e-02 5.70148349e-01
-5.80228865e-01 -8.65016103e-01 -6.68610811e-01 -1.16271722e+00
5.81124365e-01 -3.13498318e-01 1.26289165e+00 1.02184311e-01
2.05521435e-01 9.68657807e-02 -1.21145141e+00 -6.60412490e-01
-2.84445107e-01 4.76834297e-01 1.09550014e-01 -1.34780064e-01
8.60287905e-01 -1.87367901e-01 -2.97672182e-01 2.63214111e-04
-1.11626899e+00 -7.61595964e-01 4.33048397e-01 9.03124928e-01
1.59853697e-01 -2.35409662e-01 6.30614042e-01 -8.67219269e-01
1.09967661e+00 -6.03622735e-01 -2.66695917e-01 1.93104252e-01
-5.45932114e-01 1.08641997e-01 6.18415296e-01 -2.38591835e-01
-4.48383152e-01 -2.63183117e-01 -5.72944999e-01 3.37012798e-01
7.27261929e-03 6.50974691e-01 1.62563488e-01 6.12707324e-02
7.99431384e-01 -8.83777812e-02 1.07608244e-01 -2.19135508e-01
5.00947595e-01 9.77680922e-01 -1.03948630e-01 -4.04644817e-01
4.90953952e-01 4.33760494e-01 -2.63887107e-01 -9.33602214e-01
-5.28101027e-01 -1.15163215e-01 -8.56488109e-01 -9.85915661e-02
9.37605560e-01 -5.58655381e-01 -4.74289119e-01 2.20301449e-01
-1.04642951e+00 3.11395694e-02 -1.72553770e-02 4.80024993e-01
-8.01612809e-02 2.82268077e-01 -6.91013992e-01 -7.08675683e-01
-3.84417176e-01 -1.40181732e+00 7.95156598e-01 2.10087761e-01
-2.63996571e-01 -1.42345309e+00 4.59554464e-01 2.41860282e-03
5.08225918e-01 3.23575824e-01 8.45249057e-01 -1.05444598e+00
3.93166751e-01 -3.25480968e-01 -4.11226563e-02 7.08620131e-01
1.02599636e-01 5.11465013e-01 -1.08543694e+00 -2.48735160e-01
-1.88278645e-01 -4.12881196e-01 7.31672883e-01 1.37790650e-01
6.16633952e-01 -1.34730354e-01 2.17265308e-01 1.05507046e-01
1.54609919e+00 1.52687490e-01 7.78188467e-01 9.64299142e-01
3.74883294e-01 8.30260932e-01 7.05821574e-01 4.28015560e-01
3.91658783e-01 2.88845301e-01 5.73363423e-01 1.11799598e-01
3.61421376e-01 4.10243720e-01 9.41958308e-01 1.43304360e+00
-5.70181087e-02 -4.65224057e-01 -1.07078719e+00 8.87021482e-01
-1.35078132e+00 -4.62833762e-01 -2.88578063e-01 1.71169841e+00
7.03904152e-01 1.83175087e-01 8.43461677e-02 4.43158895e-01
3.88395160e-01 4.69297856e-01 3.39205682e-01 -1.50755811e+00
-4.41441089e-01 6.99742973e-01 2.88122118e-01 5.81237733e-01
-1.26599419e+00 9.80912745e-01 5.85073853e+00 5.47118425e-01
-1.25284553e+00 2.29553476e-01 4.62318212e-01 9.79552269e-02
-4.15838689e-01 -4.06141222e-01 -6.96719110e-01 3.77743721e-01
1.33995330e+00 1.77100345e-01 -3.22884582e-02 7.20451236e-01
-2.59460598e-01 9.10943449e-02 -6.99122846e-01 5.49984753e-01
5.44121265e-01 -1.21059370e+00 5.42045198e-02 -1.50951847e-01
6.74998820e-01 1.67075247e-01 4.49376255e-01 3.68708760e-01
1.53903991e-01 -1.31470847e+00 4.54521865e-01 1.68294102e-01
4.27848458e-01 -1.16861975e+00 1.46756935e+00 -3.33911240e-01
-9.15937304e-01 2.89527271e-02 -6.24401093e-01 -5.46636991e-02
-5.19268699e-02 4.23207402e-01 -6.17272973e-01 6.35260344e-01
7.81114697e-01 1.02195847e+00 -9.50925112e-01 2.90179521e-01
-1.51821986e-01 5.60912967e-01 -1.98068947e-01 -6.89718962e-01
6.79569244e-01 -5.85241020e-01 3.12876910e-01 1.55935526e+00
2.25024536e-01 -3.22152495e-01 -4.30919349e-01 3.41565728e-01
-3.39817628e-02 7.44212270e-01 -6.52425885e-01 -3.62015128e-01
2.50638008e-01 1.52692533e+00 -9.42014754e-01 -8.13195407e-02
-4.69540656e-01 8.62472117e-01 3.25718373e-01 -6.78303316e-02
-7.87938178e-01 -8.03098083e-01 7.30413318e-01 -3.69646996e-01
4.30834174e-01 -4.26211506e-01 -1.56823382e-01 -1.11714005e+00
4.41736504e-02 -9.13948774e-01 2.48376742e-01 -6.00376070e-01
-1.49353433e+00 1.24180365e+00 -2.64759421e-01 -1.16382694e+00
5.00839017e-03 -1.51018667e+00 -3.31473976e-01 7.68953323e-01
-1.65269971e+00 -1.12203741e+00 1.10994324e-01 1.64778814e-01
2.87110865e-01 -5.66843390e-01 1.11874855e+00 4.01947796e-01
-4.83425438e-01 7.82538772e-01 2.99536765e-01 4.51735467e-01
8.72086585e-01 -1.46356845e+00 8.29623267e-02 6.06052756e-01
3.31811100e-01 7.60774314e-01 6.51069343e-01 -9.81623977e-02
-1.38330257e+00 -6.30242229e-01 1.24761140e+00 -8.75287414e-01
1.05279469e+00 -3.75233650e-01 -8.17911446e-01 5.62181056e-01
8.80068362e-01 -1.29707560e-01 1.09391534e+00 2.94548333e-01
-3.67865711e-01 -7.50934780e-02 -1.11579907e+00 6.47751570e-01
1.85077086e-01 -3.55768263e-01 -8.17027152e-01 3.21346253e-01
5.68907440e-01 -1.69992894e-01 -1.16328406e+00 2.82515049e-01
5.90518594e-01 -8.99918020e-01 7.28218496e-01 -5.79389274e-01
8.92707944e-01 -2.47570723e-01 -4.63877141e-01 -1.44542694e+00
1.90635137e-02 -1.51346740e-03 9.00386870e-02 1.14981425e+00
8.21461439e-01 -8.14293742e-01 4.53128219e-01 -6.73290119e-02
-1.77115828e-01 -1.12385261e+00 -5.42592585e-01 -3.66609931e-01
7.73763895e-01 -3.74983490e-01 5.46327651e-01 1.20501244e+00
1.88327342e-01 1.43825516e-01 5.22227511e-02 1.21788338e-01
-1.77163586e-01 -2.28398085e-01 3.19581181e-01 -9.75161493e-01
3.55308086e-01 -4.93544430e-01 -6.39564633e-01 -3.38207930e-01
2.65146613e-01 -9.39011335e-01 -3.11014205e-01 -1.31997883e+00
-4.67126995e-01 -4.65150237e-01 -5.91858268e-01 4.33190525e-01
1.35474309e-01 7.01142728e-01 1.57577097e-01 -1.50407299e-01
-4.14909542e-01 6.01271689e-01 8.97809148e-01 -2.46437918e-02
7.48934876e-03 -7.52205312e-01 -6.58461690e-01 6.01168573e-01
1.15873170e+00 -3.18057716e-01 -2.65263878e-02 -5.53261518e-01
8.94627273e-01 -8.68051231e-01 -2.36310139e-01 -8.84350359e-01
-2.95052469e-01 2.28197664e-01 3.09913158e-01 -4.95683134e-01
3.60700756e-01 -8.04987311e-01 -6.28869295e-01 4.02581513e-01
-2.43221432e-01 7.82235086e-01 3.82415146e-01 -1.10149244e-02
-5.49486816e-01 -7.92407572e-01 7.03175008e-01 1.00586765e-01
-6.17365301e-01 -1.17531739e-01 -5.76579928e-01 -1.03678279e-01
8.28417897e-01 -1.72118079e-02 -3.94958049e-01 -9.85812023e-02
-6.53424561e-01 -8.25764239e-02 5.34710288e-01 6.17487669e-01
5.47138453e-01 -1.34566879e+00 -7.43903399e-01 1.57725170e-01
2.75948554e-01 -6.13717079e-01 -2.48427972e-01 8.37559462e-01
-1.21740639e+00 3.53676051e-01 -5.65780878e-01 -1.00726061e-01
-1.07260907e+00 3.14386159e-01 1.55208319e-01 -1.65446311e-01
6.63683861e-02 8.27183962e-01 -7.20172882e-01 -9.26083207e-01
-1.51141390e-01 -2.70706505e-01 -1.11437142e+00 7.75226176e-01
4.50893462e-01 3.41709346e-01 4.32406098e-01 -1.06078959e+00
-5.80186725e-01 4.66984719e-01 -5.68300523e-02 -2.82812506e-01
1.61296761e+00 2.51797974e-01 -6.21830046e-01 7.47641802e-01
1.44267535e+00 4.04435664e-01 -3.84974539e-01 3.78755927e-01
6.23637531e-03 -2.32219279e-01 -1.51436791e-01 -6.61615491e-01
-9.96868253e-01 1.00993025e+00 8.70480895e-01 6.03474259e-01
8.03458452e-01 -3.04163098e-01 6.86804354e-01 5.53509474e-01
-1.39503539e-01 -1.23505068e+00 2.16253415e-01 1.00913274e+00
8.87771606e-01 -1.37523282e+00 4.85836007e-02 5.39243920e-03
-8.96847486e-01 1.54936457e+00 4.04007733e-01 -6.50690138e-01
1.06942797e+00 7.47338608e-02 5.45111775e-01 -3.63287985e-01
-5.44160783e-01 -1.30593345e-01 3.34358633e-01 7.42662191e-01
1.09320319e+00 3.90333012e-02 -9.10173833e-01 6.27603829e-01
-6.18046463e-01 -6.08918428e-01 9.14684415e-01 9.35431540e-01
-2.67321646e-01 -1.52772272e+00 -3.12583268e-01 2.91160792e-01
-7.34947741e-01 -2.63086766e-01 -5.32984972e-01 9.73375738e-01
2.34222516e-01 9.28085923e-01 2.78937787e-01 -2.98700154e-01
1.67027399e-01 1.73608512e-01 3.26943547e-01 -4.75696623e-01
-1.00287294e+00 -4.22673434e-01 1.89816058e-01 -4.44121361e-02
-6.72623754e-01 -6.36622965e-01 -1.21225023e+00 -2.95029193e-01
-4.43878531e-01 3.17254841e-01 9.74444330e-01 6.99409783e-01
1.25642285e-01 3.66603017e-01 4.68326896e-01 -6.60284579e-01
-1.74642459e-01 -1.30425382e+00 -6.78097785e-01 3.84407699e-01
2.21478462e-01 -6.46269143e-01 -5.36711872e-01 -8.44640099e-03]
|
[11.102437019348145, 7.168318748474121]
|
a8e29e16-5b80-4f37-b6a1-6581162ecfdb
|
disentangled-generation-network-for-enlarged
|
2206.00859
| null |
https://arxiv.org/abs/2206.00859v2
|
https://arxiv.org/pdf/2206.00859v2.pdf
|
Disentangled Generation Network for Enlarged License Plate Recognition and A Unified Dataset
|
License plate recognition plays a critical role in many practical applications, but license plates of large vehicles are difficult to be recognized due to the factors of low resolution, contamination, low illumination, and occlusion, to name a few. To overcome the above factors, the transportation management department generally introduces the enlarged license plate behind the rear of a vehicle. However, enlarged license plates have high diversity as they are non-standard in position, size, and style. Furthermore, the background regions contain a variety of noisy information which greatly disturbs the recognition of license plate characters. Existing works have not studied this challenging problem. In this work, we first address the enlarged license plate recognition problem and contribute a dataset containing 9342 images, which cover most of the challenges of real scenes. However, the created data are still insufficient to train deep methods of enlarged license plate recognition, and building large-scale training data is very time-consuming and high labor cost. To handle this problem, we propose a novel task-level disentanglement generation framework based on the Disentangled Generation Network (DGNet), which disentangles the generation into the text generation and background generation in an end-to-end manner to effectively ensure diversity and integrity, for robust enlarged license plate recognition. Extensive experiments on the created dataset are conducted, and we demonstrate the effectiveness of the proposed approach in three representative text recognition frameworks.
|
['Jin Tang', 'Ruoran Jia', 'Chang Tan', 'Aihua Zheng', 'Guohao Wang', 'Xiaobin Yang', 'Chenglong Li']
|
2022-06-02
| null | null | null | null |
['license-plate-recognition']
|
['computer-vision']
|
[ 2.34461308e-01 -7.44804502e-01 7.55768940e-02 -1.04357481e-01
-7.69049466e-01 -6.73409462e-01 5.25930464e-01 -7.80022562e-01
-6.46953210e-02 6.84713721e-01 4.36188020e-02 -3.50564718e-02
1.80143178e-01 -7.23982275e-01 -5.47044516e-01 -1.08703351e+00
5.63182652e-01 1.93752557e-01 2.88903296e-01 -1.49973467e-01
4.01156157e-01 3.90042603e-01 -1.47709417e+00 2.69894749e-01
1.19458663e+00 7.67536879e-01 1.19658552e-01 2.69804537e-01
-6.05732165e-02 4.01954830e-01 -7.42843986e-01 -7.09794462e-01
4.48045284e-01 -2.73596376e-01 2.45857641e-01 5.97278178e-01
5.72819531e-01 -5.88264227e-01 -8.42165232e-01 1.29341364e+00
5.92268765e-01 7.73176178e-02 6.52891219e-01 -1.36771429e+00
-9.07365263e-01 6.95446599e-03 -6.84603035e-01 1.02635041e-01
-1.80345580e-01 2.92633444e-01 6.01673126e-01 -1.20825803e+00
3.11515838e-01 9.67784286e-01 4.89919424e-01 6.31609797e-01
-7.50098467e-01 -1.10303557e+00 5.37598804e-02 1.16223291e-01
-1.48941767e+00 -6.96128428e-01 8.48337233e-01 -4.72305566e-01
4.02790695e-01 3.25232297e-01 3.69877547e-01 1.20597470e+00
-6.27269298e-02 1.01467335e+00 9.92350161e-01 -3.82222012e-02
-1.49512231e-01 2.01142102e-01 -1.49296494e-02 5.31342328e-01
7.45261610e-01 -1.28675744e-01 -2.01219633e-01 2.33541787e-01
8.12848210e-01 3.07778895e-01 -4.80095267e-01 -2.00995967e-01
-1.07605231e+00 5.33837199e-01 -2.10112426e-02 -4.30413410e-02
9.44658443e-02 -3.61040801e-01 2.06986815e-01 -2.60849804e-01
3.88117731e-01 -5.31629771e-02 1.35358065e-01 -6.91607744e-02
-9.12545741e-01 1.61334947e-01 5.63251376e-01 1.13258016e+00
5.09378433e-01 3.61460418e-01 -3.62708941e-02 1.11952114e+00
5.11899352e-01 9.97188151e-01 3.83427143e-01 -1.94859684e-01
1.24982476e+00 5.87163985e-01 7.02827722e-02 -1.39072204e+00
5.27272336e-02 -3.58659476e-01 -1.14269173e+00 2.90324807e-01
2.79940248e-01 -1.13605388e-01 -8.89287710e-01 1.21574533e+00
-1.22311518e-01 2.49976218e-01 1.32422179e-01 1.03474820e+00
9.60797846e-01 9.76159811e-01 -2.32773364e-01 -7.20982403e-02
1.23001468e+00 -1.18526912e+00 -8.40624213e-01 -5.35151720e-01
9.68078896e-02 -9.40060437e-01 8.20952177e-01 2.72993147e-01
-7.67990708e-01 -4.31030422e-01 -1.43246162e+00 1.02805220e-01
-2.12619677e-01 6.73310816e-01 1.93292007e-01 9.47631180e-01
-5.10848641e-01 -1.61555439e-01 -3.69151086e-01 2.08363026e-01
6.63284361e-01 2.57229358e-01 -4.61274266e-01 -6.64932966e-01
-1.15594792e+00 7.73497760e-01 5.83200976e-02 7.49817193e-01
-3.19465131e-01 -1.71834037e-01 -7.89756775e-01 -6.72273710e-02
5.18183172e-01 -3.23309839e-01 6.43735945e-01 -7.30044544e-01
-1.42767394e+00 5.46962976e-01 2.58216076e-02 3.71592581e-01
7.72156715e-01 2.34727617e-02 -8.89273405e-01 -2.29585424e-01
2.60604098e-02 1.25867590e-01 8.13900530e-01 -1.33823526e+00
-5.21441400e-01 -6.55483723e-01 -2.47027084e-01 4.33914453e-01
-3.30422193e-01 -4.27823327e-03 -9.74906862e-01 -6.67003930e-01
7.05535337e-02 -9.62685049e-01 9.70015079e-02 -4.04876500e-01
-5.89770198e-01 5.75067587e-02 1.05004489e+00 -9.25743699e-01
9.99785423e-01 -2.53344607e+00 -3.52121204e-01 3.55002359e-02
3.44969928e-01 6.64209962e-01 -1.54052988e-01 1.32676512e-01
1.92005277e-01 3.23334336e-01 -1.35958821e-01 -3.69323850e-01
2.05662698e-01 -6.95810691e-02 -3.79772842e-01 4.83393341e-01
3.02862406e-01 6.79836392e-01 -4.43680465e-01 -3.83164763e-01
2.30628833e-01 4.51780587e-01 -1.30685687e-01 4.04266594e-03
1.93834201e-01 2.22938061e-02 -6.66000962e-01 9.34000015e-01
1.29976249e+00 2.14932203e-01 -2.19608486e-01 -2.99819499e-01
-2.58485615e-01 -2.84636468e-01 -1.27253938e+00 9.82940972e-01
-1.18303806e-01 8.78844440e-01 -2.60002445e-02 -5.63332200e-01
1.13563788e+00 1.01817861e-01 1.04099393e-01 -8.66587460e-01
6.00981675e-02 3.49871933e-01 -9.19859409e-02 -8.08170080e-01
7.42859125e-01 -8.16260129e-02 -1.12733617e-01 1.22488700e-02
-5.01291215e-01 1.07357532e-01 3.05026501e-01 -8.90195519e-02
8.44434321e-01 -5.40246740e-02 -2.38112375e-01 2.95547962e-01
6.37512207e-01 -6.98042288e-02 7.87082791e-01 3.59910846e-01
-3.42407286e-01 9.60108876e-01 4.17067438e-01 -2.34995127e-01
-1.09643662e+00 -9.49491799e-01 5.72400279e-02 5.24174035e-01
5.41588426e-01 1.50644153e-01 -6.32322133e-01 -4.75302756e-01
-2.06011817e-01 4.18236613e-01 -1.95451707e-01 -1.76535830e-01
-7.52328992e-01 -1.03280139e+00 9.48523819e-01 4.68251884e-01
1.07949352e+00 -6.80076122e-01 3.10534060e-01 -8.58637616e-02
-4.52071488e-01 -1.43032300e+00 -8.83644283e-01 -6.87999487e-01
-4.85264033e-01 -1.08281422e+00 -8.30795527e-01 -9.93444920e-01
8.76763403e-01 8.08076262e-01 6.22285485e-01 -6.03755936e-02
-1.44680172e-01 -3.43239784e-01 -6.32229894e-02 -3.32028478e-01
-3.32402885e-01 -2.83687234e-01 -5.49597703e-02 4.55248088e-01
3.36041391e-01 -1.22461081e-01 -5.50511658e-01 7.44938970e-01
-1.13185394e+00 2.97857434e-01 9.69322562e-01 7.79578507e-01
2.24172547e-01 5.46246648e-01 3.98489684e-01 -5.05686462e-01
7.17655599e-01 -2.83177495e-01 -8.84713411e-01 3.32693070e-01
-2.72829235e-01 -4.27807063e-01 7.52911687e-01 -3.32185239e-01
-1.17696989e+00 3.06124166e-02 -6.02038316e-02 -4.52827364e-01
-1.82682753e-01 3.34979028e-01 -9.47425246e-01 -1.26584455e-01
7.82039091e-02 7.54511178e-01 1.26446217e-01 -2.22724691e-01
-3.47027816e-02 9.36620951e-01 6.24238491e-01 -2.15814590e-01
1.10753202e+00 1.27920732e-01 -1.26390561e-01 -8.82522941e-01
-1.59387425e-01 -3.69349599e-01 -2.79169440e-01 -3.55398923e-01
6.00207090e-01 -9.10676718e-01 -5.76643646e-01 1.10549569e+00
-1.19602728e+00 2.64831990e-01 4.74914074e-01 4.62709248e-01
8.50974843e-02 6.46677971e-01 -3.73766422e-01 -8.25003743e-01
-3.83245423e-02 -1.45837271e+00 1.06554377e+00 5.18901885e-01
7.29021847e-01 -5.41605949e-01 -1.60757825e-01 9.20724869e-01
4.06191200e-01 1.27951846e-01 8.36080194e-01 -6.67477071e-01
-1.11849427e+00 -5.79817951e-01 -6.14707649e-01 5.20566702e-01
1.37302145e-01 2.02205926e-01 -7.06830025e-01 -1.53236799e-02
-1.48298651e-01 -1.10649079e-01 9.11536396e-01 2.81162877e-02
7.27265954e-01 -2.62716830e-01 -1.39498577e-01 5.44018745e-01
1.37166548e+00 3.92692327e-01 1.16704500e+00 3.45833927e-01
1.00763977e+00 5.54879189e-01 5.00476897e-01 3.96903694e-01
3.58489513e-01 6.95417881e-01 3.07428062e-01 -6.64413869e-02
-5.69238402e-02 -9.16479230e-02 5.63869774e-01 8.82460952e-01
-1.99313417e-01 -7.55433857e-01 -8.09290409e-01 5.57215624e-02
-1.61213505e+00 -1.17412531e+00 -3.41545254e-01 2.13958073e+00
3.89287859e-01 1.74868293e-02 -1.17751479e-01 1.21908084e-01
1.23075724e+00 2.80546695e-01 -5.42884350e-01 3.11537027e-01
-4.10754979e-01 -5.91778219e-01 5.40311933e-01 4.63819727e-02
-1.11371589e+00 7.59230018e-01 5.44025087e+00 1.05162203e+00
-1.26387310e+00 -3.13469440e-01 5.69298685e-01 1.21660732e-01
-1.71820417e-01 -3.32014471e-01 -1.04158425e+00 9.63021398e-01
2.13396803e-01 -4.89996970e-02 3.87755901e-01 5.37320316e-01
1.37326449e-01 1.90751761e-01 -7.27509856e-01 1.38185847e+00
5.84863722e-01 -1.14543641e+00 9.11239162e-02 2.40336239e-01
6.34042561e-01 -9.52531025e-02 4.05301601e-01 4.05117124e-01
-2.11702585e-01 -9.61877704e-01 5.62020898e-01 3.77571642e-01
8.24702024e-01 -7.61490166e-01 8.24213743e-01 4.86922294e-01
-1.00915468e+00 1.03980377e-01 -4.79019046e-01 3.54894251e-01
-6.30634204e-02 4.08617467e-01 -6.92086935e-01 5.71570158e-01
6.07082881e-02 5.68792164e-01 -5.25663972e-01 1.18001068e+00
-8.62260088e-02 4.44890738e-01 -8.77100825e-02 -1.57885566e-01
1.61281452e-01 -5.86184382e-01 5.93261600e-01 1.11440265e+00
6.30726099e-01 1.24487408e-01 8.74337479e-02 9.58474398e-01
-2.48141527e-01 -5.51661570e-03 -6.34448290e-01 -2.08347321e-01
4.06134307e-01 1.18011701e+00 -7.33518660e-01 -3.83556113e-02
-7.65147924e-01 9.53629553e-01 -2.58276820e-01 5.05654752e-01
-1.07695198e+00 -5.52368045e-01 5.24031341e-01 -1.53240278e-01
3.05126607e-01 -3.17392379e-01 -9.87267569e-02 -1.60525131e+00
4.84462380e-01 -1.06207836e+00 -2.75791995e-02 -7.25239694e-01
-1.32447457e+00 5.24340034e-01 -3.88754725e-01 -1.72099531e+00
1.09701395e-01 -8.85655403e-01 -5.22121429e-01 9.29128647e-01
-1.34737122e+00 -1.22526622e+00 -6.69016182e-01 5.32690942e-01
9.10539985e-01 -5.99605620e-01 4.02702779e-01 7.45307565e-01
-1.26116335e+00 6.76987052e-01 6.82262242e-01 6.02974236e-01
6.40774548e-01 -6.75097942e-01 4.40687925e-01 1.19972432e+00
-1.58880964e-01 4.40104246e-01 3.36180717e-01 -6.81317389e-01
-1.73462224e+00 -1.20823216e+00 4.66725945e-01 -2.52324969e-01
3.62539172e-01 -5.75989306e-01 -8.48249793e-01 3.35273594e-01
-1.35407016e-01 1.46364644e-01 7.21802771e-01 -3.28829676e-01
-3.51395637e-01 -2.86079317e-01 -8.61118376e-01 7.88151741e-01
7.32940912e-01 -2.34388903e-01 -4.70672488e-01 9.42365602e-02
2.95827031e-01 -4.27617401e-01 -2.50026643e-01 2.82547176e-01
7.18011737e-01 -9.21610415e-01 8.95098805e-01 -1.27326116e-01
5.32634735e-01 -5.45944810e-01 -2.07548186e-01 -1.21143055e+00
-2.84870267e-01 -2.52890408e-01 3.33366811e-01 1.63647890e+00
3.41322184e-01 -8.56077552e-01 9.18132961e-01 9.62541997e-01
-4.08330083e-01 -5.41816115e-01 -7.94283271e-01 -8.07480752e-01
-2.76597470e-01 -1.83934852e-01 7.23040640e-01 8.42396617e-01
-4.29264277e-01 2.74107367e-01 -7.59354353e-01 4.72720116e-01
6.74806356e-01 1.15630314e-01 8.77509475e-01 -9.20361817e-01
-5.52813942e-03 -5.80175161e-01 -8.11076880e-01 -1.09972882e+00
7.05088004e-02 -5.71881056e-01 3.74525070e-01 -1.48563159e+00
4.19434130e-01 -4.18114483e-01 -1.16476312e-01 4.85767163e-02
-4.08946067e-01 4.78739887e-01 3.08787227e-01 5.18773794e-01
-5.73323965e-01 6.71946943e-01 1.29233706e+00 -5.07629454e-01
1.90578494e-02 1.66309059e-01 -7.61011541e-01 6.33689880e-01
6.47147298e-01 -2.07802787e-01 -4.23197985e-01 -6.73179746e-01
2.94510610e-02 1.34327024e-01 2.67498165e-01 -8.84810507e-01
4.31956977e-01 -1.36487469e-01 5.74549973e-01 -7.88617313e-01
4.35132176e-01 -8.03157508e-01 4.98978533e-02 2.94905659e-02
2.04287797e-01 -9.67842489e-02 2.78673559e-01 6.68397069e-01
-4.49444026e-01 -1.27381429e-01 4.71355617e-01 1.59077987e-01
-1.00661588e+00 5.78456700e-01 -5.16695857e-01 6.60124421e-02
1.00479293e+00 -6.04551971e-01 -7.56608367e-01 -1.14190042e-01
4.41267192e-02 1.26632750e-01 4.03817952e-01 6.31843746e-01
8.42795253e-01 -1.46311331e+00 -1.01403439e+00 5.11177838e-01
1.90593809e-01 -1.35603193e-02 6.30684197e-01 6.13607109e-01
-6.61879599e-01 3.21205586e-01 -2.71902233e-01 -3.14883620e-01
-1.35490072e+00 2.20961004e-01 1.68312564e-01 -1.44201338e-01
-3.76779318e-01 4.88502920e-01 3.99462551e-01 -2.95586526e-01
-1.13444484e-03 6.87314123e-02 -2.56686568e-01 -4.81140092e-02
6.35574937e-01 5.12413383e-01 1.05583370e-01 -9.86061513e-01
-2.35746548e-01 7.41144776e-01 -1.22061409e-01 7.45066851e-02
1.07224631e+00 -9.23834816e-02 3.86540918e-03 4.32220697e-02
1.14804566e+00 4.41113502e-01 -1.30011845e+00 -7.05565587e-02
-5.38094759e-01 -9.60236788e-01 -1.77960813e-01 -6.40333951e-01
-1.17733335e+00 8.31223369e-01 3.20050508e-01 -1.32693620e-02
9.98122156e-01 -6.70625925e-01 9.47834492e-01 5.34517527e-01
3.23937088e-01 -1.01441324e+00 5.21625318e-02 4.48536187e-01
8.66393924e-01 -1.30810916e+00 -5.75880818e-02 -7.18076646e-01
-8.24884951e-01 1.24231935e+00 9.18420672e-01 1.08985037e-01
3.31649035e-02 3.13838154e-01 1.08246192e-01 1.93768561e-01
-2.38532230e-01 1.26048505e-01 3.27739149e-01 5.16707897e-01
1.49974078e-01 -5.58388792e-02 3.58063094e-02 8.91810000e-01
1.80662304e-01 -2.08535627e-01 5.84967971e-01 6.63132608e-01
-2.89867252e-01 -1.02731621e+00 -7.50703633e-01 3.52591068e-01
-2.96987951e-01 -1.42745331e-01 -3.77351224e-01 1.00974214e+00
2.74318099e-01 9.96499300e-01 -6.72124028e-02 -6.22640133e-01
4.49650854e-01 -1.77281478e-03 2.81497508e-01 -1.54264927e-01
2.31744483e-01 4.74594906e-02 4.77633215e-02 1.48861930e-01
-1.62622094e-01 -5.81844032e-01 -9.74588752e-01 -4.77010936e-01
-5.95411062e-01 -1.66045904e-01 6.75233066e-01 9.79471862e-01
2.91556001e-01 3.25975180e-01 9.10165668e-01 -8.02156568e-01
-6.14491284e-01 -7.54231513e-01 -8.09395611e-01 5.14328837e-01
1.53236791e-01 -6.56860769e-01 -2.60453641e-01 9.87142622e-02]
|
[9.856074333190918, -4.917244911193848]
|
47676900-f4fd-4d1d-a82d-a923c4ea2e5f
|
don-t-discard-fixed-window-audio-segmentation
|
2210.13363
| null |
https://arxiv.org/abs/2210.13363v1
|
https://arxiv.org/pdf/2210.13363v1.pdf
|
Don't Discard Fixed-Window Audio Segmentation in Speech-to-Text Translation
|
For real-life applications, it is crucial that end-to-end spoken language translation models perform well on continuous audio, without relying on human-supplied segmentation. For online spoken language translation, where models need to start translating before the full utterance is spoken, most previous work has ignored the segmentation problem. In this paper, we compare various methods for improving models' robustness towards segmentation errors and different segmentation strategies in both offline and online settings and report results on translation quality, flicker and delay. Our findings on five different language pairs show that a simple fixed-window audio segmentation can perform surprisingly well given the right conditions.
|
['Barry Haddow', 'Chantal Amrhein']
|
2022-10-24
| null | null | null | null |
['speech-to-text-translation']
|
['natural-language-processing']
|
[ 2.22552791e-01 -2.78369011e-03 -2.30676606e-01 -5.66536427e-01
-1.66172862e+00 -8.80452394e-01 4.60187376e-01 5.62100336e-02
-3.83033395e-01 5.98514438e-01 2.33200967e-01 -7.50406921e-01
4.84256208e-01 -4.86958597e-04 -5.46022952e-01 -6.20967299e-02
1.86303243e-01 7.31761217e-01 3.93493772e-01 -3.22213858e-01
1.40842021e-01 6.29190728e-02 -1.04528522e+00 1.66084111e-01
1.06579661e+00 5.39887190e-01 3.83904763e-02 1.12255204e+00
-2.39232630e-01 2.67481983e-01 -1.06672406e+00 -1.70813873e-01
2.16835812e-01 -1.02671576e+00 -8.70442986e-01 6.24861494e-02
2.97856152e-01 -3.02098602e-01 -1.77383840e-01 7.71604836e-01
1.02245975e+00 -6.58508167e-02 2.23686814e-01 -1.18894207e+00
-3.89951617e-01 1.04844904e+00 2.84046475e-02 3.86948735e-01
9.86373425e-01 1.97077528e-01 8.29556465e-01 -8.21541250e-01
6.12024724e-01 1.13470745e+00 6.30558789e-01 3.13205957e-01
-1.39645052e+00 -6.32598400e-01 9.17258486e-02 -1.41482577e-01
-1.45160675e+00 -1.19279778e+00 2.73823887e-01 -1.99396431e-01
1.22696877e+00 4.02987212e-01 4.54568565e-01 9.62253869e-01
2.52882361e-01 6.04948640e-01 1.05384982e+00 -7.76106119e-01
1.20687455e-01 3.37742269e-01 4.02730927e-02 1.36178330e-01
-5.62907755e-01 3.94978002e-02 -9.12459850e-01 -1.72233000e-01
2.25180626e-01 -8.12636733e-01 -2.96995252e-01 4.54044372e-01
-1.38535440e+00 5.36716878e-01 -3.91388297e-01 3.12478334e-01
-1.11958213e-01 -6.88110366e-02 4.92000103e-01 9.01627541e-01
7.72782445e-01 4.52818543e-01 -4.61572081e-01 -9.26338673e-01
-1.46418405e+00 -4.63253073e-02 1.13166499e+00 1.27594662e+00
3.71649474e-01 1.57516316e-01 -2.47505814e-01 9.88128901e-01
1.32005781e-01 7.95394123e-01 3.87294441e-01 -9.43577647e-01
5.61573982e-01 -2.42620289e-01 2.68260032e-01 -4.98537809e-01
-2.60510385e-01 -2.60586381e-01 -1.22834750e-01 -4.94722694e-01
5.54888666e-01 -5.52159965e-01 -8.59352350e-01 1.48943102e+00
1.06899917e-01 -8.87695998e-02 -4.00933325e-02 9.64692235e-01
4.69581544e-01 9.78388965e-01 -3.31633210e-01 -6.60541892e-01
8.91717255e-01 -1.05172896e+00 -1.19350922e+00 -3.88248801e-01
6.00968838e-01 -1.56715643e+00 1.45485723e+00 5.35827518e-01
-1.34006727e+00 -3.15051556e-01 -9.55578744e-01 -3.40097621e-02
6.48048520e-02 -8.89759436e-02 1.28308132e-01 8.88616204e-01
-1.48600376e+00 2.79194027e-01 -7.58691311e-01 -6.45681143e-01
-4.76280987e-01 3.50060970e-01 5.99511527e-02 -1.74406134e-02
-1.41255999e+00 8.93126488e-01 -2.36060515e-01 2.06593741e-02
-5.48855841e-01 -3.39025885e-01 -5.80727220e-01 -1.01618938e-01
3.09272975e-01 -1.97902828e-01 2.10819411e+00 -1.21594262e+00
-2.13189197e+00 5.05465209e-01 -5.87705553e-01 -2.44287997e-01
7.21179783e-01 -3.46042067e-01 -4.48232800e-01 1.80769265e-01
-5.98375909e-02 6.50149524e-01 8.28002393e-01 -1.11185110e+00
-4.94966954e-01 1.10497460e-01 -4.40802753e-01 4.02485281e-01
-2.26943381e-02 6.64707959e-01 -6.20106518e-01 -5.15843451e-01
1.68878466e-01 -9.36381876e-01 1.08308382e-01 -3.98730636e-01
-4.12568837e-01 3.24980952e-02 6.70903265e-01 -1.13928926e+00
1.49707508e+00 -2.21866393e+00 4.39319611e-02 2.97386438e-01
-5.12057543e-01 -1.58073788e-03 -1.98774114e-01 7.64456809e-01
4.09626991e-01 2.78649539e-01 5.37855141e-02 -6.24577224e-01
2.06347659e-01 7.84253255e-02 -4.31472331e-01 4.16143686e-01
-4.05908525e-02 7.75442481e-01 -7.42054999e-01 -5.99663615e-01
2.23223842e-03 4.28767264e-01 -3.47646743e-01 2.24722460e-01
-1.97896779e-01 5.18023431e-01 2.38378942e-01 6.10022604e-01
1.75378919e-01 3.47807616e-01 -3.31753772e-03 6.15662515e-01
-3.04192245e-01 8.99123132e-01 -9.44719613e-01 1.69010258e+00
-6.45797014e-01 1.11099112e+00 5.44825912e-01 -3.02202731e-01
7.11500823e-01 7.17473447e-01 2.44591221e-01 -9.52243686e-01
1.69434175e-01 5.71381867e-01 1.01826452e-01 -2.82452494e-01
6.55035377e-01 -3.04722451e-02 -5.90713285e-02 9.53199148e-01
-4.36137281e-02 -6.65932059e-01 5.14959432e-02 2.03882158e-01
6.70268476e-01 -2.19189107e-01 -1.82228789e-01 -1.09034225e-01
-1.36786535e-01 3.05828601e-01 1.76270738e-01 8.41071248e-01
-4.59573776e-01 9.16397095e-01 2.47042850e-01 3.05131674e-01
-1.19267011e+00 -8.20403099e-01 1.73597723e-01 1.56917894e+00
3.56902629e-02 -4.18778718e-01 -1.26521087e+00 -5.12240715e-02
-5.39273024e-01 9.90930080e-01 2.91358113e-01 -1.21960334e-01
-4.59739119e-01 -1.54398561e-01 1.06438243e+00 1.18488967e-01
3.74922603e-02 -6.68833017e-01 -3.87932181e-01 6.30054295e-01
-7.77501702e-01 -1.20442665e+00 -1.26655221e+00 7.39484876e-02
-6.95073664e-01 -1.86752543e-01 -7.48377502e-01 -8.09476078e-01
4.63009030e-02 4.72634971e-01 1.06004739e+00 -5.35928942e-02
3.64120342e-02 4.62360501e-01 -5.86650550e-01 -4.67162102e-01
-8.77068162e-01 3.15172702e-01 1.73154324e-01 -3.07425261e-01
2.60222673e-01 -3.17135453e-01 -1.34308055e-01 6.83614671e-01
-5.88133633e-01 -6.76953569e-02 2.73665190e-01 7.45009422e-01
2.43524522e-01 -2.97893614e-01 6.28750265e-01 -3.40580106e-01
1.09818387e+00 -6.13220297e-02 -3.24734926e-01 3.02235633e-01
-6.00755930e-01 -3.11510921e-01 5.42372346e-01 -6.83830082e-01
-6.56656563e-01 -6.56106323e-02 -2.61770159e-01 -6.67781532e-02
-1.30002961e-01 6.45593166e-01 -6.59128949e-02 1.83704183e-01
7.32887328e-01 4.37359691e-01 2.06778660e-01 -1.54626787e-01
2.18688577e-01 1.27626216e+00 2.90985644e-01 -3.77943963e-01
3.55098605e-01 -2.73198277e-01 -9.58454311e-01 -1.11376858e+00
-1.52037576e-01 -4.89304602e-01 -4.39177901e-01 -3.25510055e-01
5.31278074e-01 -9.72513735e-01 -2.22080410e-01 5.25221527e-01
-1.36513829e+00 -8.03169489e-01 4.14876193e-02 6.23047113e-01
-6.41987860e-01 2.47113883e-01 -7.42767572e-01 -9.92776215e-01
-1.57348081e-01 -1.32831490e+00 1.28441870e+00 -2.75689494e-02
-7.69905686e-01 -7.90334463e-01 3.31398621e-02 3.08914363e-01
6.86205089e-01 -6.19816184e-01 4.83789653e-01 -7.60180891e-01
-3.66434723e-01 -2.04565436e-01 1.89916700e-01 9.22613516e-02
1.03588149e-01 4.47990417e-01 -1.05018187e+00 -3.27770948e-01
-2.87802935e-01 -3.15664232e-01 3.08284730e-01 4.46978778e-01
1.12820871e-01 -4.08593029e-01 4.17622291e-02 1.37507394e-01
6.56956017e-01 3.50919425e-01 3.43502045e-01 -7.11563155e-02
2.31065318e-01 4.74010706e-01 7.31927156e-01 3.85844141e-01
2.89406210e-01 8.33201766e-01 -4.34248894e-01 -1.74117982e-01
3.69359814e-02 -3.52642834e-01 9.21139956e-01 1.40823030e+00
4.45426822e-01 -6.69030190e-01 -1.07626736e+00 5.83667338e-01
-1.63411486e+00 -6.00127280e-01 8.55720118e-02 2.19800615e+00
1.14887452e+00 3.63652021e-01 4.82617021e-01 2.69037813e-01
7.25094140e-01 -2.93816123e-02 -1.99916542e-01 -9.58616853e-01
-2.90849861e-02 2.12131858e-01 4.92947787e-01 1.16514874e+00
-6.53302550e-01 1.22816837e+00 8.35395718e+00 7.61556923e-01
-1.41366613e+00 3.88396800e-01 5.26090562e-01 -2.69417584e-01
-4.02639300e-01 1.35287428e-02 -6.06437027e-01 3.78761679e-01
1.64890289e+00 -2.98262268e-01 7.38511443e-01 3.05845261e-01
1.01881957e+00 -4.60397393e-01 -1.29611075e+00 8.45373034e-01
-1.10738873e-01 -7.65143514e-01 -3.66630793e-01 -1.93640754e-01
4.25705135e-01 1.58086850e-03 3.68720703e-02 2.37770140e-01
9.87916961e-02 -9.63823497e-01 1.08456004e+00 1.38400733e-01
1.03239870e+00 -6.77941680e-01 5.85279047e-01 5.52296281e-01
-9.04703557e-01 3.31313521e-01 1.66322917e-01 -2.24498600e-01
5.41388273e-01 3.31525654e-01 -1.32718408e+00 3.09750468e-01
2.34030485e-01 1.43070877e-01 -3.45549077e-01 8.55209947e-01
-2.43636742e-02 1.34230602e+00 -7.78991103e-01 -1.32411048e-01
2.29322806e-01 -4.03940603e-02 5.05048990e-01 1.63609457e+00
4.80681837e-01 6.60012439e-02 3.26829642e-01 4.25547510e-01
2.49307454e-01 3.30224067e-01 -4.66656446e-01 -4.26177889e-01
8.01710665e-01 4.35294479e-01 -9.08497274e-01 -3.03645968e-01
-2.81122595e-01 1.29305589e+00 -4.31832343e-01 7.11570382e-01
-7.99638808e-01 -5.72657406e-01 4.44218040e-01 1.09713182e-01
-7.46381581e-02 -7.16948926e-01 -5.82405567e-01 -8.41095507e-01
3.71271297e-02 -1.32562149e+00 -1.16017237e-01 -7.02047586e-01
-6.06841385e-01 5.54933727e-01 -1.87222183e-01 -1.00831473e+00
-6.43166423e-01 9.99195650e-02 -4.49774027e-01 9.35985267e-01
-1.00108874e+00 -8.13935399e-01 3.47551823e-01 3.52214962e-01
1.08834791e+00 1.00151367e-01 7.93663561e-01 4.81711000e-01
-3.16999495e-01 1.04172063e+00 2.20444277e-01 -1.30739838e-01
1.26711261e+00 -8.98359120e-01 6.25086248e-01 9.53685522e-01
4.28710520e-01 4.77535248e-01 1.29375470e+00 -5.52225769e-01
-1.50881088e+00 -5.86768031e-01 1.47452438e+00 -3.64495903e-01
5.03770173e-01 -6.88464463e-01 -8.44879150e-01 5.99900484e-01
7.48759866e-01 -6.23915732e-01 6.78947091e-01 5.74881621e-02
-4.67524081e-02 4.82771657e-02 -8.53453755e-01 5.69000363e-01
7.88676679e-01 -9.36175287e-01 -3.32916260e-01 4.13851142e-01
1.17586684e+00 -6.43879712e-01 -5.88461280e-01 -2.59414613e-01
5.59811831e-01 -6.76422298e-01 3.88914317e-01 -3.15538675e-01
-4.69411202e-02 -6.17416911e-02 -2.00788364e-01 -1.60665560e+00
2.03531474e-01 -1.42632794e+00 3.78441453e-01 1.22014916e+00
9.23207760e-01 -6.83040023e-01 2.20559686e-01 8.62911046e-01
-3.97785366e-01 -1.54035747e-01 -1.07174218e+00 -8.68915498e-01
5.66983642e-03 -6.36246443e-01 2.83129305e-01 7.44552851e-01
4.89739627e-01 6.00181341e-01 -3.18573713e-01 1.74701914e-01
-5.38207591e-02 -2.18639538e-01 7.89608598e-01 -5.98626673e-01
-2.38676399e-01 -5.40492952e-01 -4.77064140e-02 -1.29072213e+00
-4.75977957e-02 -3.65451634e-01 7.13778436e-01 -1.33584428e+00
-2.84019113e-01 -1.94240525e-01 2.29009509e-01 2.63654262e-01
-1.88135996e-01 1.92966789e-01 2.99819887e-01 7.09960759e-02
-6.05204403e-01 4.27297890e-01 9.50383663e-01 -1.95439443e-01
-6.50360286e-01 1.50906742e-01 -3.82276952e-01 4.94047664e-02
7.17749953e-01 -4.04525459e-01 -5.04081070e-01 -6.88310802e-01
5.91382794e-02 4.15373266e-01 -1.58080116e-01 -7.75448382e-01
4.72862959e-01 -1.80419162e-01 -2.79447943e-01 -4.55681145e-01
3.22522849e-01 -5.77558279e-01 1.67481527e-01 1.67055428e-01
-5.50112903e-01 2.79659331e-01 3.99927109e-01 1.19989119e-01
-6.13803804e-01 4.51233797e-02 5.94884813e-01 2.07272843e-01
-1.86199918e-01 -1.12350553e-01 -1.08262217e+00 2.34915107e-01
6.55354917e-01 -4.10710961e-01 1.68098867e-01 -1.10619128e+00
-9.13370669e-01 3.47565562e-01 5.36517918e-01 4.29281145e-01
4.41196680e-01 -1.07176709e+00 -7.72116303e-01 2.26996198e-01
-1.56205833e-01 -2.89224595e-01 -2.54041940e-01 1.08616114e+00
-5.24724305e-01 5.07751167e-01 3.49098772e-01 -1.00426769e+00
-1.42002285e+00 4.65456806e-02 3.25181842e-01 3.15212190e-01
-3.02402973e-01 9.11324501e-01 -7.04295695e-01 -4.59486008e-01
7.17045367e-01 -5.33480108e-01 6.12429976e-01 1.14599213e-01
4.26725894e-01 1.93123221e-01 4.00903314e-01 -6.43766224e-01
-2.01970607e-01 2.16601610e-01 3.37887853e-02 -1.22505021e+00
5.31855464e-01 -6.74152493e-01 1.96979582e-01 1.05761588e+00
7.41963208e-01 1.69553936e-01 -9.63684499e-01 -2.35844597e-01
6.27327785e-02 -3.90301466e-01 1.02087129e-02 -8.81859541e-01
-3.90835255e-01 1.01608634e+00 4.00085062e-01 4.01649296e-01
9.27313089e-01 -2.03237325e-01 1.15474725e+00 3.37145180e-01
5.93023539e-01 -1.53431380e+00 -1.80320039e-01 9.57422018e-01
7.20806479e-01 -1.13447928e+00 -4.25781995e-01 -2.85319984e-01
-7.40700662e-01 8.95427823e-01 1.70754805e-01 4.86070037e-01
3.77261877e-01 4.69429314e-01 8.20189416e-01 5.15830636e-01
-8.58686030e-01 9.88172814e-02 4.54591447e-03 5.22240877e-01
7.32527494e-01 2.20811635e-01 -2.57224560e-01 4.00335461e-01
-6.19476914e-01 -1.37148559e-01 6.31366193e-01 1.03373325e+00
-6.96442246e-01 -1.12516201e+00 -7.72657514e-01 9.99701098e-02
-4.30241525e-01 -2.73670375e-01 -1.16101027e+00 5.02721965e-01
-4.30187523e-01 1.62510931e+00 6.68415055e-02 -4.54876810e-01
4.33398485e-01 7.24273324e-01 4.07999218e-01 -7.13927090e-01
-8.32197785e-01 8.53987813e-01 5.27070940e-01 -4.55269516e-01
-2.38338143e-01 -8.61638963e-01 -1.33839536e+00 -4.98612791e-01
-6.44033909e-01 3.46666008e-01 7.14990377e-01 1.04517877e+00
3.96943629e-01 2.59972960e-01 7.22073257e-01 -8.99589360e-01
-6.77886307e-01 -1.20600569e+00 -2.96742201e-01 -1.73118621e-01
6.73807383e-01 1.53411711e-02 -3.62387955e-01 2.66993314e-01]
|
[14.598464012145996, 6.995684623718262]
|
accb72df-001d-4fb0-910f-24a570b12634
|
self-supervised-learning-for-biologically
|
2303.02370
| null |
https://arxiv.org/abs/2303.02370v2
|
https://arxiv.org/pdf/2303.02370v2.pdf
|
Self-Supervised Learning for Place Representation Generalization across Appearance Changes
|
Visual place recognition is a key to unlocking spatial navigation for animals, humans and robots. While state-of-the-art approaches are trained in a supervised manner and therefore hardly capture the information needed for generalizing to unusual conditions, we argue that self-supervised learning may help abstracting the place representation so that it can be foreseen, irrespective of the conditions. More precisely, in this paper, we investigate learning features that are robust to appearance modifications while sensitive to geometric transformations in a self-supervised manner. This dual-purpose training is made possible by combining the two self-supervision main paradigms, \textit{i.e.} contrastive and predictive learning. Our results on standard benchmarks reveal that jointly learning such appearance-robust and geometry-sensitive image descriptors leads to competitive visual place recognition results across adverse seasonal and illumination conditions, without requiring any human-annotated labels.
|
['Djamila Aouada', 'Vincent Gaudillière', 'Mohamed Adel Musallam']
|
2023-03-04
| null | null | null | null |
['visual-place-recognition']
|
['computer-vision']
|
[ 4.51063454e-01 8.00367165e-03 -3.17655683e-01 -6.05809510e-01
-2.20376655e-01 -8.25819075e-01 9.85445976e-01 3.32641572e-01
-5.02163768e-01 7.28797674e-01 -1.54720142e-01 -1.70976356e-01
-2.17892915e-01 -5.21699190e-01 -9.55979943e-01 -8.29307377e-01
-2.79462636e-01 3.99718046e-01 3.02262187e-01 -3.95195127e-01
4.98556554e-01 8.77123535e-01 -2.06337595e+00 -1.12952761e-01
7.86694646e-01 7.64191687e-01 3.28255683e-01 3.75953406e-01
1.30695254e-01 5.80437362e-01 -1.18488885e-01 1.86565995e-01
5.29306769e-01 -3.09736431e-01 -8.24744165e-01 2.25279152e-01
6.59612536e-01 5.70730641e-02 -1.18566684e-01 8.75623941e-01
2.68099099e-01 3.20119321e-01 9.30332601e-01 -1.15153182e+00
-6.70385540e-01 -1.27213374e-01 -3.06750327e-01 6.68131337e-02
5.69601893e-01 2.56390601e-01 8.40498924e-01 -6.62377298e-01
1.06781578e+00 8.72566819e-01 5.92265129e-01 3.46306175e-01
-1.62346280e+00 -3.27104986e-01 2.96169430e-01 2.89726973e-01
-1.26362538e+00 -7.88637280e-01 1.04224873e+00 -4.92565423e-01
8.06181848e-01 1.67714208e-01 4.15725976e-01 1.05494297e+00
1.75526425e-01 6.00468397e-01 1.55420244e+00 -5.81228018e-01
4.47774887e-01 6.09050579e-02 -4.67483133e-01 6.60555720e-01
3.40705335e-01 4.25253540e-01 -8.24836075e-01 2.34838277e-01
7.59871483e-01 1.52908161e-01 -2.10720152e-01 -1.19726515e+00
-1.40426803e+00 3.40297967e-01 8.63207757e-01 3.72619301e-01
-3.55225146e-01 -1.08007498e-01 1.71412721e-01 2.72424489e-01
3.90009642e-01 7.24080145e-01 -4.04843897e-01 1.45876437e-01
-9.12088871e-01 1.07197657e-01 4.40774590e-01 1.10949433e+00
1.37156308e+00 1.29711756e-03 6.38704896e-02 6.02708697e-01
3.25370908e-01 7.70764768e-01 3.54863018e-01 -7.63616443e-01
5.76932281e-02 5.52028656e-01 1.08572111e-01 -1.10276461e+00
-7.23438680e-01 -4.16752875e-01 -7.26893365e-01 5.30185163e-01
4.35610801e-01 2.56156445e-01 -1.19924438e+00 1.61701250e+00
3.55859965e-01 -3.36475596e-02 -3.28688025e-02 6.80841088e-01
4.45085645e-01 2.89662629e-01 2.71202531e-02 4.56233658e-02
1.04772019e+00 -1.04704595e+00 -3.02974135e-01 -7.96652317e-01
6.41782761e-01 -5.80214858e-01 8.47090840e-01 3.14333290e-01
-4.34014976e-01 -4.81063336e-01 -1.25946283e+00 8.14727172e-02
-8.50494921e-01 -1.26726791e-01 8.11419010e-01 5.41871369e-01
-1.21330392e+00 7.18561292e-01 -6.61394477e-01 -9.45089877e-01
3.26813877e-01 4.36554462e-01 -1.07150388e+00 4.92724031e-03
-7.31951892e-01 9.79582787e-01 4.80631679e-01 1.75055951e-01
-9.16785121e-01 -2.20607251e-01 -1.14950991e+00 -3.88354033e-01
7.58016780e-02 -4.95591402e-01 8.84876430e-01 -1.12143540e+00
-1.45441425e+00 1.46534085e+00 -3.24003041e-01 -5.54920137e-01
4.80908215e-01 4.96861897e-02 -2.49441013e-01 1.92259774e-01
2.77924806e-01 9.73891973e-01 1.06846309e+00 -1.64999080e+00
-5.36413550e-01 -5.49566388e-01 -4.14053611e-02 2.65025705e-01
1.37156853e-02 -4.65388417e-01 -1.74083561e-01 -4.30074602e-01
4.95332330e-01 -1.22700012e+00 -3.62584591e-01 2.99068928e-01
-3.71247858e-01 2.00883776e-01 6.55676425e-01 -3.11861038e-01
5.34192920e-01 -2.03194380e+00 1.21212475e-01 3.80377442e-01
-6.99546263e-02 1.17892034e-01 -1.10252485e-01 4.62729305e-01
-7.81277046e-02 -2.70797402e-01 -3.58832330e-01 -2.73710430e-01
5.18633798e-02 5.40724874e-01 -2.62091964e-01 9.44734871e-01
1.59589157e-01 7.57733166e-01 -1.07423568e+00 -2.10084975e-01
6.11377954e-01 3.61066431e-01 -4.95905548e-01 3.27771008e-01
-1.29394874e-01 1.01067984e+00 -2.74378121e-01 8.39035392e-01
6.16826475e-01 -1.17206136e-02 7.98217878e-02 1.16378479e-01
-3.80966306e-01 2.35492766e-01 -9.92986143e-01 1.98250377e+00
-5.16359389e-01 6.06070220e-01 -1.74228966e-01 -1.00144029e+00
1.23466754e+00 -2.77337700e-01 2.97454149e-01 -1.11419046e+00
-1.14135258e-01 3.32795113e-01 -3.48698497e-01 -2.31289908e-01
3.00795883e-01 -8.24229494e-02 2.51466092e-02 8.75209048e-02
2.45982215e-01 -1.22647539e-01 -1.61898900e-02 -3.91283154e-01
7.33137071e-01 6.47053003e-01 5.36407232e-01 -5.19524276e-01
5.39233804e-01 7.44836926e-02 3.87549579e-01 8.21012378e-01
-3.77406150e-01 6.62109196e-01 -9.57284123e-03 -6.58223033e-01
-9.85361457e-01 -1.09925866e+00 -2.94904739e-01 1.32822084e+00
5.21550775e-01 -6.62630200e-02 -4.21035767e-01 -4.64484692e-01
2.58461535e-02 4.84959155e-01 -8.61981511e-01 -2.65784740e-01
-2.84247756e-01 -4.23846990e-01 3.12654674e-01 4.93789166e-01
4.32491124e-01 -1.15981781e+00 -1.03633642e+00 -2.44815685e-02
1.70177102e-01 -9.58883405e-01 7.21511543e-02 8.35074365e-01
-7.24318624e-01 -1.04589272e+00 -7.02070117e-01 -9.23155069e-01
8.36258769e-01 4.63572860e-01 8.18131745e-01 -2.09799558e-02
-1.87866762e-01 5.24394751e-01 -4.32071954e-01 -2.49072075e-01
8.51675272e-02 1.70534208e-01 2.24994153e-01 4.51099277e-02
7.74016604e-02 -9.19428110e-01 -6.27689183e-01 3.85187417e-01
-8.46511662e-01 1.74516551e-02 7.49872863e-01 8.99111092e-01
8.14507365e-01 -4.29229558e-01 1.18027478e-01 -8.02616537e-01
-1.17782190e-01 -3.11778516e-01 -5.71041763e-01 3.29250962e-01
-8.20667863e-01 4.67705846e-01 6.15056396e-01 -1.53157756e-01
-9.12423193e-01 6.18198633e-01 3.75676416e-02 -6.44895772e-05
-8.65268528e-01 1.56396553e-01 -2.02434704e-01 -7.27346957e-01
9.71532464e-01 5.83371878e-01 -1.13986969e-01 -3.00567985e-01
4.94344205e-01 2.81446636e-01 6.91487670e-01 -6.00839972e-01
1.10670292e+00 9.36490357e-01 4.17137057e-01 -1.07661629e+00
-7.47074068e-01 -8.11959326e-01 -1.33563304e+00 -1.62431583e-01
8.46735537e-01 -7.77406216e-01 -3.66545916e-01 4.08178955e-01
-9.12747979e-01 -3.54754508e-01 -1.54700890e-01 1.68416023e-01
-9.07451570e-01 2.64525592e-01 1.65969759e-01 -6.75973237e-01
1.99752256e-01 -8.32093120e-01 1.20837700e+00 3.75701010e-01
-5.28644547e-02 -1.13161564e+00 3.42062622e-01 1.39633670e-01
4.23254192e-01 4.94198740e-01 4.15434092e-01 -7.88743258e-01
-7.17923999e-01 -2.18683884e-01 -2.26015933e-02 5.85142821e-02
1.98581383e-01 -2.85603762e-01 -1.40274096e+00 -2.79682547e-01
-3.40930045e-01 -3.88154656e-01 8.17021489e-01 -3.49106104e-03
8.20069313e-01 -1.96202978e-01 -3.45296890e-01 1.03259003e+00
1.29804003e+00 8.79300535e-02 7.09316492e-01 9.50056970e-01
6.60446346e-01 7.13308990e-01 5.57695746e-01 2.54000515e-01
5.62726676e-01 7.37900138e-01 7.43918180e-01 -3.28735143e-01
2.34027326e-01 -5.43263793e-01 9.34107900e-02 1.91762239e-01
-2.06836060e-01 1.03120774e-01 -1.05365717e+00 5.90051889e-01
-1.83332491e+00 -8.80167305e-01 4.06250581e-02 2.34167886e+00
5.63208461e-01 3.26251946e-02 -1.47293404e-01 1.89758033e-01
4.04771924e-01 4.96630520e-01 -5.42529821e-01 -2.33680978e-01
-3.77084613e-01 2.22222637e-02 8.57654810e-01 2.92158425e-01
-1.52763736e+00 1.08624768e+00 6.54476452e+00 3.62256020e-01
-1.43537855e+00 -2.29462266e-01 3.87313724e-01 5.61402917e-01
-2.70706832e-01 3.07762980e-01 -4.18699473e-01 1.36305427e-03
5.44618249e-01 2.16166392e-01 5.42356849e-01 1.05323207e+00
-1.86489955e-01 -3.06578130e-01 -1.19685125e+00 1.09467304e+00
3.61921936e-01 -1.19878280e+00 -6.61357343e-02 4.03274074e-02
7.92291880e-01 2.07378976e-02 2.40155712e-01 2.52452213e-02
-1.39525205e-01 -1.20647383e+00 9.05198514e-01 5.69544852e-01
6.48997128e-01 -3.26716781e-01 6.13573134e-01 3.56516510e-01
-1.21629381e+00 3.60695049e-02 -2.30814338e-01 -2.79337943e-01
-3.54949199e-02 8.07731077e-02 -8.13168943e-01 7.14714885e-01
6.93746924e-01 1.09648204e+00 -1.12417352e+00 1.26796365e+00
-4.36200291e-01 1.53539509e-01 -4.15099591e-01 1.86854720e-01
3.67409468e-01 -7.05646351e-02 4.96408075e-01 1.06683385e+00
1.01399854e-01 -3.52334201e-01 2.32704982e-01 7.12064981e-01
4.14944887e-01 2.59402931e-01 -1.09297276e+00 2.41849259e-01
2.83166170e-01 9.99341011e-01 -9.91806448e-01 1.38333261e-01
-1.17462456e-01 1.34643078e+00 4.53003138e-01 4.08962429e-01
-3.36565942e-01 -2.67601222e-01 4.39796180e-01 1.96665108e-01
4.67550159e-01 -4.01916921e-01 -4.03561056e-01 -1.10885632e+00
5.02770999e-03 -3.39341015e-01 9.34186056e-02 -7.63290048e-01
-1.04639053e+00 4.63195413e-01 -3.61164287e-02 -1.42190886e+00
-2.90972799e-01 -9.69054759e-01 -4.52083617e-01 5.53836167e-01
-2.04762912e+00 -1.62221372e+00 -3.70250076e-01 5.96302867e-01
1.99024245e-01 8.77167750e-03 1.08245921e+00 -3.20688225e-02
-1.19529702e-01 4.75783497e-01 2.68555313e-01 -1.45974845e-01
9.58270431e-01 -1.42320013e+00 9.35418978e-02 8.70656729e-01
5.04490793e-01 8.27469051e-01 9.89566386e-01 -2.52228051e-01
-1.36529016e+00 -9.20180738e-01 9.46438789e-01 -4.97656137e-01
4.47010964e-01 -4.51954991e-01 -7.97142565e-01 5.70260644e-01
3.90504822e-02 4.73299533e-01 5.61470687e-01 1.64940372e-01
-6.95785344e-01 -2.46560633e-01 -1.09572923e+00 6.19139552e-01
1.26112843e+00 -8.66043925e-01 -6.61992788e-01 2.63019145e-01
1.93647936e-01 -4.05573487e-01 -5.55621862e-01 4.66005623e-01
4.28918213e-01 -1.25644898e+00 1.15568411e+00 -4.41657513e-01
7.19665661e-02 -6.45137131e-01 -2.28950411e-01 -1.19633627e+00
-4.08360064e-01 -6.00117385e-01 3.32267344e-01 9.59837258e-01
3.95408303e-01 -7.76178837e-01 6.29579008e-01 2.51810402e-01
-1.36143684e-01 -2.76011884e-01 -1.06447077e+00 -1.00959051e+00
-6.65348396e-02 -2.49438643e-01 3.64692509e-01 1.07966721e+00
6.96393009e-03 2.02746317e-02 -4.98734623e-01 3.70368868e-01
7.31762469e-01 1.97774753e-01 9.12972271e-01 -1.42274678e+00
1.42465100e-01 -3.70953530e-01 -9.93244171e-01 -1.03027260e+00
4.35093522e-01 -9.47977304e-01 4.21323061e-01 -1.42858779e+00
-1.11732356e-01 -4.57811236e-01 -3.82115990e-01 7.20981836e-01
2.25094795e-01 3.17395002e-01 -2.31317729e-02 3.30466539e-01
-8.87585282e-01 6.56893432e-01 1.07388842e+00 -1.79990008e-01
-1.29095569e-01 -1.79022893e-01 -4.65329587e-01 7.01435268e-01
7.27716923e-01 -2.93633997e-01 -3.47766250e-01 -2.56352782e-01
1.02904826e-01 -5.75659633e-01 7.35370219e-01 -1.25223017e+00
4.51986223e-01 -4.22884375e-01 4.69157100e-01 -3.23688537e-01
1.15016960e-01 -9.95738983e-01 -3.31195188e-03 3.16551596e-01
-1.97067708e-01 -1.02572121e-01 2.39994809e-01 8.15227449e-01
-1.81635693e-01 2.08671689e-02 7.81759322e-01 -8.71829689e-02
-1.33270586e+00 7.59235770e-02 -2.49170661e-01 -1.48599312e-01
1.02610850e+00 -5.96952081e-01 -2.12231338e-01 -2.73847193e-01
-5.54289043e-01 -4.33575884e-02 1.08244717e+00 5.18310845e-01
4.33885872e-01 -1.30002403e+00 -1.62015870e-01 5.59540391e-01
8.67564142e-01 7.22392574e-02 2.31227785e-01 7.29253709e-01
-8.80507231e-01 5.96998334e-01 -6.79287970e-01 -9.27474022e-01
-8.49110663e-01 7.44223535e-01 3.21197480e-01 1.44532442e-01
-5.50412476e-01 6.09790444e-01 2.67303586e-01 -8.52888346e-01
2.73952603e-01 -2.68824935e-01 -2.09447816e-01 -8.16937461e-02
2.34718725e-01 5.17052375e-02 1.11904018e-01 -1.04046810e+00
-6.65829301e-01 7.65190780e-01 2.63135582e-01 3.94674577e-02
1.42351103e+00 -3.88071865e-01 -7.03681111e-02 6.74864292e-01
9.44854736e-01 -1.12203449e-01 -1.48071837e+00 -1.75190613e-01
4.07795638e-01 -3.67158353e-01 -2.12560952e-01 -6.98997855e-01
-6.63436830e-01 8.47335815e-01 8.07876348e-01 -1.09093539e-01
9.73730206e-01 -8.37824941e-02 2.54580408e-01 8.01879525e-01
8.57801795e-01 -1.08394873e+00 -1.33923888e-01 6.12295985e-01
7.88661182e-01 -1.53443348e+00 4.56330292e-02 -2.08652645e-01
-5.45126498e-01 1.05217659e+00 4.86133754e-01 -1.48948655e-01
6.26924157e-01 -1.93081841e-01 2.10585430e-01 -1.41852140e-01
-2.63493955e-01 -4.99770701e-01 7.48244584e-01 1.16767752e+00
2.66212612e-01 -6.64747972e-03 2.94300199e-01 3.11918259e-02
-7.66407028e-02 -3.85512024e-01 1.98865429e-01 1.31455314e+00
-6.28443241e-01 -9.59238529e-01 -4.26958621e-01 -9.31318924e-02
2.31447116e-01 2.50732154e-02 -5.19651175e-01 6.29877687e-01
3.37317139e-01 5.65558493e-01 -1.45473525e-01 -3.72287154e-01
3.43643010e-01 1.91003382e-01 5.89590371e-01 -6.01506174e-01
-2.21442327e-01 -2.21190140e-01 -1.02490947e-01 -6.79191530e-01
-7.78245687e-01 -7.37219751e-01 -1.07172203e+00 1.12192184e-01
-4.01335172e-02 -2.46376038e-01 5.77380478e-01 1.12387824e+00
2.89318174e-01 6.97206557e-02 5.03811896e-01 -1.51677656e+00
-1.72663555e-01 -5.30193686e-01 -6.17637336e-01 5.95513344e-01
6.98174417e-01 -1.03537154e+00 -3.90933096e-01 2.71692038e-01]
|
[7.744194030761719, -1.9312554597854614]
|
85a2d29a-e012-4529-84d1-1c28fa3d2973
|
investigating-audio-visual-and-text-fusion
|
1805.00705
| null |
http://arxiv.org/abs/1805.00705v2
|
http://arxiv.org/pdf/1805.00705v2.pdf
|
Investigating Audio, Visual, and Text Fusion Methods for End-to-End Automatic Personality Prediction
|
We propose a tri-modal architecture to predict Big Five personality trait
scores from video clips with different channels for audio, text, and video
data. For each channel, stacked Convolutional Neural Networks are employed. The
channels are fused both on decision-level and by concatenating their respective
fully connected layers. It is shown that a multimodal fusion approach
outperforms each single modality channel, with an improvement of 9.4\% over the
best individual modality (video). Full backpropagation is also shown to be
better than a linear combination of modalities, meaning complex interactions
between modalities can be leveraged to build better models. Furthermore, we can
see the prediction relevance of each modality for each trait. The described
model can be used to increase the emotional intelligence of virtual agents.
|
['Pascale Fung', 'Dario Bertero', 'Onno Kampman', 'Elham J. Barezi']
|
2018-05-02
| null | null | null | null |
['emotional-intelligence']
|
['natural-language-processing']
|
[ 1.56500623e-01 3.78546596e-01 5.31859696e-02 -5.68628848e-01
-5.43631732e-01 -2.97954798e-01 5.85273921e-01 1.30054966e-01
-3.44204634e-01 5.70788145e-01 3.76937628e-01 3.06977183e-01
-2.21282154e-01 -4.77950513e-01 -6.44976377e-01 -5.78792095e-01
-2.39197776e-01 1.05155539e-02 -2.76485056e-01 -3.67967159e-01
-3.49768735e-02 1.48344025e-01 -1.67556906e+00 8.64272654e-01
4.09063250e-01 1.59903598e+00 -1.61391735e-01 8.02542627e-01
6.17038384e-02 1.09949064e+00 -5.30601799e-01 -7.88929760e-01
3.09344213e-02 -9.86569002e-03 -5.91083884e-01 1.72059074e-01
4.78931308e-01 -3.35717440e-01 -3.46428066e-01 8.71577621e-01
4.36580181e-01 1.24117784e-01 4.37378317e-01 -1.26290452e+00
-5.33233821e-01 8.30663919e-01 -4.12428796e-01 -3.26468915e-01
7.16769159e-01 3.96822840e-02 1.06129754e+00 -6.84294999e-01
4.11084980e-01 1.25830173e+00 5.97405314e-01 5.09278953e-01
-1.13984907e+00 -4.86765653e-01 1.93285555e-01 4.15446430e-01
-7.36343205e-01 -4.41498250e-01 9.18325961e-01 -3.55070531e-01
1.15965605e+00 1.13359042e-01 7.17777252e-01 1.50175583e+00
2.17625782e-01 8.44044209e-01 1.35921562e+00 -7.98677281e-02
-8.03683326e-02 4.63482231e-01 1.04276188e-01 8.37549329e-01
-3.06021988e-01 5.61016165e-02 -1.00009739e+00 -9.01691392e-02
3.79082590e-01 3.43284719e-02 9.08504799e-02 9.17264000e-02
-1.21211898e+00 6.69041574e-01 4.65973645e-01 2.47589603e-01
-7.85764098e-01 1.58575550e-01 5.55429518e-01 5.83491087e-01
2.51257151e-01 6.82309151e-01 -3.99956524e-01 -4.07615662e-01
-7.29128242e-01 1.53436069e-03 7.89474010e-01 4.73393351e-01
5.93046904e-01 1.96899429e-01 -2.47038022e-01 9.75806713e-01
2.04802439e-01 3.20745558e-01 2.76588440e-01 -1.25357175e+00
2.13595226e-01 7.29692519e-01 -8.30133855e-02 -1.06093085e+00
-7.40839422e-01 -3.69409114e-01 -7.62011290e-01 3.93052936e-01
2.16151386e-01 -6.77788854e-01 -7.42141008e-01 1.80656648e+00
-2.44892851e-01 1.54289052e-01 2.78925180e-01 7.58339643e-01
1.20959437e+00 6.29467905e-01 1.78669229e-01 -1.54693186e-01
1.22379398e+00 -9.32819664e-01 -7.64994383e-01 -4.42574948e-01
1.68511331e-01 -2.34793618e-01 6.79645896e-01 7.32782960e-01
-1.57266808e+00 -6.25734866e-01 -1.09605694e+00 1.12492889e-01
-4.26925749e-01 1.23243049e-01 8.89255404e-01 6.68480396e-01
-1.26218903e+00 6.15673244e-01 -5.43458164e-01 -2.50701278e-01
3.57131898e-01 7.50609279e-01 -6.05318248e-01 8.30662176e-02
-1.27622902e+00 8.03341627e-01 2.86219627e-01 4.94876467e-02
-5.95656633e-01 -4.64082301e-01 -9.13119972e-01 4.12004590e-01
-5.35155162e-02 -8.31072867e-01 8.93803120e-01 -1.57683432e+00
-1.83544779e+00 6.05387449e-01 -1.54449418e-01 -5.06630957e-01
2.19918549e-01 4.85098734e-03 -6.27402604e-01 4.45677370e-01
-5.42642236e-01 1.04211175e+00 9.57765460e-01 -1.18856561e+00
-6.52615368e-01 -4.09632742e-01 3.01041543e-01 3.44611317e-01
-9.51505959e-01 9.83393937e-02 -3.45796257e-01 -6.23475425e-02
-1.58988759e-01 -8.20866883e-01 -7.08791241e-02 -6.33974552e-01
-2.25183904e-01 -7.82664269e-02 5.24390817e-01 -9.20247495e-01
1.08668566e+00 -2.18289828e+00 6.32960916e-01 2.26319730e-01
2.46277422e-01 -1.95817441e-01 -3.16759557e-01 2.96887577e-01
4.42082882e-02 -9.26499739e-02 2.74819672e-01 -7.75849044e-01
3.27161819e-01 -1.06675178e-03 1.78552672e-01 -1.15348302e-01
3.95867616e-01 8.84841025e-01 -2.74853647e-01 -1.05296038e-01
1.85513198e-01 9.34210598e-01 -4.34386075e-01 2.02333853e-02
8.06536898e-02 4.55303818e-01 1.07428834e-01 7.24847496e-01
3.69126141e-01 -1.80778176e-01 2.35612258e-01 -1.25440881e-01
3.62768292e-01 -1.47407278e-01 -7.21766353e-01 1.50918758e+00
-4.50300276e-01 7.35834360e-01 2.87127972e-01 -6.70443356e-01
1.10077500e+00 5.03401220e-01 7.00687706e-01 -7.23728776e-01
3.41272116e-01 -2.59792656e-01 -4.79186662e-02 -6.48608148e-01
7.35166788e-01 -2.50452131e-01 -4.44119513e-01 3.27745155e-02
5.09442747e-01 3.40704352e-01 8.33176449e-02 1.48546129e-01
1.06723130e+00 -1.10004440e-01 -2.67585605e-01 4.30831850e-01
6.05528891e-01 -4.42701757e-01 4.30043787e-01 5.90123415e-01
-3.45243007e-01 3.09604615e-01 9.91213202e-01 -2.36164346e-01
-7.86622822e-01 -5.02474606e-01 1.32671624e-01 1.65221894e+00
-1.65126786e-01 -3.59966308e-01 -6.19833171e-01 -5.95252633e-01
3.76071855e-02 4.32188839e-01 -8.89222980e-01 -2.30234981e-01
-6.40083626e-02 -5.50372660e-01 6.31654620e-01 7.71477818e-01
4.57820922e-01 -1.11250675e+00 -5.42467535e-01 -8.67040530e-02
-1.34469643e-01 -1.38136351e+00 3.59054834e-01 3.42891812e-01
-5.71944118e-01 -4.97205257e-01 -5.29583871e-01 -5.31628072e-01
1.72236755e-01 -1.77999258e-01 9.29479837e-01 -1.74581632e-01
5.04016340e-01 8.06230605e-01 -3.60239357e-01 -3.57922614e-01
-2.36764222e-01 -1.01259291e-01 1.87269703e-01 3.91343832e-01
3.98715585e-01 -5.55505395e-01 -4.05857980e-01 -3.86038683e-02
-5.35907805e-01 1.14309072e-01 7.44767249e-01 1.01980662e+00
1.21458225e-01 3.71372029e-02 6.78762734e-01 -3.55061442e-01
8.65866065e-01 -6.03094578e-01 3.50371040e-02 2.17665657e-01
-3.06562603e-01 -2.49545977e-01 4.73949969e-01 -4.63438898e-01
-1.34014988e+00 1.80355161e-01 -1.85387179e-01 -5.66458285e-01
-6.89561427e-01 8.44483018e-01 -1.19410656e-01 -1.86815217e-01
2.29006305e-01 -1.63972348e-01 8.02593231e-02 -1.93928048e-01
2.68284529e-01 7.42978275e-01 6.29319429e-01 -3.18092108e-01
-4.15446870e-02 2.63758689e-01 -9.49412733e-02 -7.40154326e-01
-4.51252669e-01 -2.01750129e-01 -4.52996910e-01 -7.13908076e-01
1.01831424e+00 -9.99128282e-01 -1.29060280e+00 5.69895029e-01
-7.51141310e-01 -2.12079391e-01 1.64296001e-01 5.99094689e-01
-5.55586040e-01 3.73319685e-02 -1.02281880e+00 -8.21030259e-01
-2.14945331e-01 -1.10786068e+00 9.01220262e-01 6.06726944e-01
-1.91195950e-01 -1.21217334e+00 -3.13210517e-01 8.45487833e-01
3.37630689e-01 1.28484845e-01 6.67506576e-01 -7.92290390e-01
1.53626904e-01 -5.40154278e-01 -2.73424089e-01 3.65658283e-01
-4.72193241e-01 6.26827357e-03 -1.41154218e+00 -2.06988797e-01
-2.57429242e-01 -8.14392567e-01 1.15927160e+00 5.05515635e-01
9.29197431e-01 -9.71241519e-02 1.89384222e-01 4.58788395e-01
1.07951224e+00 1.33484885e-01 5.49538255e-01 1.71825081e-01
6.59025252e-01 7.22815514e-01 3.15978706e-01 7.07493663e-01
5.52861810e-01 3.06086630e-01 5.84356606e-01 -1.03540972e-01
1.66982904e-01 1.18442096e-01 8.45027447e-01 6.35629654e-01
-4.81628418e-01 -7.19430670e-02 -7.65495121e-01 1.11798555e-01
-1.97436047e+00 -1.29271209e+00 3.59399691e-02 1.87297869e+00
2.45490283e-01 1.40316561e-01 4.34650987e-01 -4.54131030e-02
4.23877597e-01 8.81166235e-02 -5.33456385e-01 -7.27301002e-01
-5.21815419e-01 -7.71846995e-02 2.07496718e-01 3.01107019e-01
-1.23239851e+00 7.70576775e-01 7.19953394e+00 3.14004093e-01
-1.29991639e+00 3.36329676e-02 6.80831730e-01 -4.99746889e-01
-1.31605044e-01 -4.91213441e-01 -4.44039524e-01 3.51946712e-01
1.25502145e+00 3.07793438e-01 6.26315355e-01 5.60040057e-01
-1.76689290e-02 -2.37842008e-01 -9.74490583e-01 1.16196251e+00
3.70712638e-01 -9.33887839e-01 -3.59116644e-01 9.21459571e-02
6.64324999e-01 -3.50235738e-02 3.64444196e-01 6.87354803e-01
1.95568964e-01 -1.13589931e+00 5.76386809e-01 1.02329862e+00
2.86938339e-01 -9.73094761e-01 1.05479074e+00 7.31476992e-02
-7.69976914e-01 -7.97530413e-01 -1.52278945e-01 -4.35327083e-01
4.86513078e-02 9.28426534e-02 -5.51274836e-01 4.49721336e-01
8.79933476e-01 7.38111496e-01 -8.46127152e-01 7.69203246e-01
1.80961341e-01 4.42120314e-01 -1.37118772e-01 -5.60885947e-03
1.54357597e-01 -2.90932879e-02 4.18326795e-01 1.33714509e+00
4.24748451e-01 -8.41346607e-02 1.35361046e-01 4.26735967e-01
-6.31815866e-02 -6.07779585e-02 -5.80003500e-01 -3.97401661e-01
1.47647783e-01 1.65879917e+00 -3.74530822e-01 -4.45561826e-01
-7.43611157e-01 1.15682125e+00 3.35642457e-01 5.05210817e-01
-9.17353392e-01 -2.34626025e-01 6.06634676e-01 -4.99530971e-01
2.94640899e-01 3.75955291e-02 -7.35777974e-01 -1.01044226e+00
-3.22368383e-01 -8.18184316e-01 5.95734715e-01 -1.01878405e+00
-1.29856825e+00 5.94947278e-01 -3.07992786e-01 -1.01900852e+00
-4.37713623e-01 -6.58697665e-01 -4.30378437e-01 5.79589963e-01
-1.10090041e+00 -1.42507100e+00 -4.65901583e-01 6.76225722e-01
1.06591642e-01 -5.22140741e-01 1.06220365e+00 2.27310404e-01
-6.98575854e-01 5.91274202e-01 -1.99352890e-01 4.09493744e-02
8.14794958e-01 -1.24617100e+00 -5.58812737e-01 5.05136728e-01
-2.62490034e-01 2.23460153e-01 6.28848493e-01 -5.28878927e-01
-1.56965375e+00 -5.78257501e-01 6.56359673e-01 -1.82239726e-01
6.35216951e-01 -7.55955186e-03 -7.49517679e-01 5.03183186e-01
9.57174659e-01 -4.93264049e-01 1.19363666e+00 6.07443333e-01
-4.99431282e-01 -1.36936486e-01 -1.13738942e+00 3.64871264e-01
5.32842219e-01 -5.59516728e-01 -2.81662881e-01 -1.64704725e-01
5.29651046e-01 -3.17293145e-02 -1.42031264e+00 3.45123738e-01
8.47547829e-01 -1.41254663e+00 9.31057274e-01 -6.84197962e-01
9.76088166e-01 2.83125073e-01 -3.81464541e-01 -1.53911364e+00
-5.63880861e-01 -3.10689390e-01 -4.24509376e-01 1.08692336e+00
5.46153903e-01 -4.04078186e-01 6.90294743e-01 1.06804359e+00
-5.01684025e-02 -7.40303099e-01 -5.54931045e-01 -2.99758673e-01
-1.02161814e-03 -6.55960679e-01 5.78290880e-01 1.06705046e+00
9.32409465e-01 6.29022300e-01 -8.38364065e-01 6.12348057e-02
2.86695510e-01 -1.40443116e-01 4.34980214e-01 -1.24638891e+00
-4.48528200e-01 -7.93668270e-01 -5.84426820e-01 -4.75023597e-01
3.10277790e-01 -6.65075362e-01 -4.17123675e-01 -1.55245388e+00
3.34417492e-01 1.37346357e-01 -7.79626906e-01 7.38013148e-01
-1.37885764e-01 6.73966229e-01 5.23117363e-01 -3.67881626e-01
-7.49081194e-01 4.58725423e-01 1.13517678e+00 -3.53166878e-01
-4.11287844e-01 -3.09806228e-01 -8.51013660e-01 8.66435111e-01
1.02018142e+00 2.92571664e-01 -1.54484436e-01 -2.67631531e-01
5.14083982e-01 7.54381537e-01 5.72649002e-01 -1.27314305e+00
2.57238686e-01 5.68841472e-02 8.37906659e-01 -2.59286582e-01
1.17947090e+00 -7.71958649e-01 4.55462821e-02 -2.78146192e-02
-6.16947234e-01 -2.78947890e-01 2.96221852e-01 3.22063565e-01
-4.05045807e-01 8.89575854e-02 3.59853119e-01 3.67351621e-02
-8.30170870e-01 -7.15331212e-02 -4.54308301e-01 -7.82875299e-01
9.57596600e-01 -3.82008553e-01 -2.18211457e-01 -9.81757462e-01
-1.27230644e+00 2.64206052e-01 9.66988504e-02 4.96012777e-01
7.33053565e-01 -1.47099328e+00 -4.78975773e-01 1.81933030e-01
-4.44350438e-03 -1.01502883e+00 7.93446720e-01 1.36477590e+00
2.94301599e-01 1.88473672e-01 -9.09238636e-01 -4.16537523e-01
-1.55997002e+00 4.08745468e-01 3.05348366e-01 -4.20789458e-02
-3.13354172e-02 9.01260793e-01 -2.06319854e-01 -2.65930027e-01
6.44249737e-01 2.60931164e-01 -8.72943878e-01 4.98348325e-01
6.02453887e-01 4.33110505e-01 -2.98879910e-02 -8.33455920e-01
-1.21936677e-02 2.01403052e-01 1.77038148e-01 -4.04506505e-01
1.64834690e+00 -2.24854916e-01 -5.47308885e-02 6.46877885e-01
9.78515506e-01 -2.04853103e-01 -1.36207736e+00 -8.02209377e-02
-3.47288489e-01 -1.48497373e-01 2.54073411e-01 -1.26607680e+00
-1.22441292e+00 1.00565457e+00 5.81538975e-01 4.63405937e-01
1.35948110e+00 -1.51028305e-01 5.50794125e-01 5.15772521e-01
1.14245817e-01 -1.32821214e+00 1.46491006e-01 5.69684505e-01
7.69331813e-01 -1.44788277e+00 -3.77554864e-01 1.78838111e-02
-1.32656431e+00 1.30499184e+00 7.14116454e-01 1.22954212e-01
4.83498216e-01 3.47606122e-01 5.65061197e-02 -2.18496010e-01
-1.07163751e+00 -2.65934557e-01 4.65903193e-01 4.73779023e-01
5.61178565e-01 3.35111648e-01 4.33562733e-02 1.22523153e+00
-7.08946288e-02 -1.81259084e-02 4.15097296e-01 6.62588239e-01
-3.58348668e-01 -8.38553727e-01 -5.15394688e-01 6.84119940e-01
-5.28020084e-01 1.40193760e-01 -7.45135725e-01 2.25220427e-01
2.25643277e-01 1.31415033e+00 6.27228245e-02 -1.21805620e+00
1.45642310e-01 5.72669446e-01 5.47521055e-01 5.25648333e-02
-1.04895663e+00 3.02556843e-01 5.41260183e-01 -6.80547774e-01
-5.11624277e-01 -7.79944658e-01 -9.56503749e-01 -5.56063473e-01
1.59849837e-01 -2.98113316e-01 6.39526129e-01 7.76159763e-01
4.29575503e-01 6.72278881e-01 5.37211537e-01 -1.23250866e+00
-9.19585675e-02 -1.08959138e+00 -4.32014823e-01 4.54387486e-01
1.07292660e-01 -6.47096455e-01 -1.34927228e-01 5.68644330e-02]
|
[13.26174259185791, 5.143737316131592]
|
f6dd11fa-ce5b-4326-b91a-fb307bedf07f
|
deepgum-learning-deep-robust-regression-with
|
1808.09211
| null |
http://arxiv.org/abs/1808.09211v1
|
http://arxiv.org/pdf/1808.09211v1.pdf
|
DeepGUM: Learning Deep Robust Regression with a Gaussian-Uniform Mixture Model
|
In this paper, we address the problem of how to robustly train a ConvNet for
regression, or deep robust regression. Traditionally, deep regression employs
the L2 loss function, known to be sensitive to outliers, i.e. samples that
either lie at an abnormal distance away from the majority of the training
samples, or that correspond to wrongly annotated targets. This means that,
during back-propagation, outliers may bias the training process due to the high
magnitude of their gradient. In this paper, we propose DeepGUM: a deep
regression model that is robust to outliers thanks to the use of a
Gaussian-uniform mixture model. We derive an optimization algorithm that
alternates between the unsupervised detection of outliers using
expectation-maximization, and the supervised training with cleaned samples
using stochastic gradient descent. DeepGUM is able to adapt to a continuously
evolving outlier distribution, avoiding to manually impose any threshold on the
proportion of outliers in the training set. Extensive experimental evaluations
on four different tasks (facial and fashion landmark detection, age and head
pose estimation) lead us to conclude that our novel robust technique provides
reliability in the presence of various types of noise and protection against a
high percentage of outliers.
|
['Xavier Alameda-Pineda', 'Stéphane Lathuilière', 'Radu Horaud', 'Pablo Mesejo']
|
2018-08-28
|
deepgum-learning-deep-robust-regression-with-1
|
http://openaccess.thecvf.com/content_ECCV_2018/html/Stephane_Lathuiliere_DeepGUM_Learning_Deep_ECCV_2018_paper.html
|
http://openaccess.thecvf.com/content_ECCV_2018/papers/Stephane_Lathuiliere_DeepGUM_Learning_Deep_ECCV_2018_paper.pdf
|
eccv-2018-9
|
['head-pose-estimation']
|
['computer-vision']
|
[-1.31188065e-01 6.29329830e-02 2.77612478e-01 -6.68967664e-01
-7.02533185e-01 1.73240583e-02 4.93473470e-01 2.16766074e-01
-5.70581675e-01 5.84157646e-01 -3.87243293e-02 2.23485619e-01
9.41616148e-02 -3.77712905e-01 -9.38843191e-01 -9.52103853e-01
-6.83073476e-02 5.33525229e-01 1.02680475e-01 -8.69244058e-03
1.67024344e-01 7.80157387e-01 -1.66440153e+00 -2.82449245e-01
7.78758645e-01 9.80206549e-01 -5.83300352e-01 3.93808454e-01
1.90748468e-01 4.12078947e-01 -9.77998793e-01 -4.07062203e-01
2.23820582e-01 -2.95199275e-01 -8.30470473e-02 1.90492898e-01
7.45699823e-01 -1.78892776e-01 8.54464397e-02 1.05053914e+00
7.80306518e-01 2.73258477e-01 8.23657513e-01 -1.41786611e+00
-2.81767070e-01 2.92610914e-01 -6.47304237e-01 5.05029112e-02
2.70150691e-01 1.89926438e-02 2.69069731e-01 -1.04775763e+00
4.34896678e-01 1.12553263e+00 9.36146736e-01 7.56325841e-01
-1.19265628e+00 -6.32359982e-01 3.36954087e-01 -3.09729446e-02
-1.58345282e+00 -8.71114433e-01 7.01399148e-01 -4.49395239e-01
3.31444710e-01 1.74253300e-01 2.10809052e-01 1.44747329e+00
2.16108188e-01 5.63506484e-01 7.75377750e-01 -2.80337781e-01
5.31087458e-01 1.34745494e-01 -3.51537131e-02 5.71513653e-01
3.98323208e-01 8.69019330e-02 -5.38175642e-01 -4.15552497e-01
2.21745074e-01 -2.03936011e-01 -2.33715042e-01 -4.74951893e-01
-3.47621530e-01 8.27142477e-01 2.97133178e-01 3.35460633e-01
-4.35908437e-01 1.92771554e-01 4.48034555e-01 3.20895940e-01
8.00086141e-01 4.71407212e-02 -3.57679337e-01 2.03577846e-01
-1.18682647e+00 2.21039087e-01 7.23374963e-01 5.61317265e-01
4.61958021e-01 3.03107798e-01 -7.25217015e-02 8.98653567e-01
6.00645542e-01 2.55780965e-01 7.73037732e-01 -5.47513008e-01
1.65029436e-01 4.06924218e-01 -4.32272479e-02 -1.04040074e+00
-5.65948248e-01 -3.91646713e-01 -8.38919997e-01 5.84680021e-01
8.25334489e-01 -2.81171113e-01 -1.14834356e+00 1.78641260e+00
5.54407835e-01 2.99597770e-01 -2.15884760e-01 7.94480920e-01
6.28669977e-01 7.60323331e-02 -6.37124479e-03 -2.14655638e-01
6.22537673e-01 -3.52677554e-01 -5.28967202e-01 -1.15179017e-01
3.73965472e-01 -5.94399452e-01 7.65368879e-01 5.31134248e-01
-9.82598960e-01 -3.97795916e-01 -9.41027343e-01 3.05140018e-01
-4.13355350e-01 1.81689207e-02 1.32809073e-01 8.49803507e-01
-1.00696194e+00 8.81772995e-01 -8.18514645e-01 -2.67239928e-01
3.75181198e-01 6.09194636e-01 -5.93601167e-01 1.26413340e-02
-6.07623100e-01 7.24843621e-01 7.75301680e-02 6.51053846e-01
-5.08872330e-01 -4.30758357e-01 -8.94078374e-01 -2.13163391e-01
1.66270405e-01 -3.14291090e-01 8.33755910e-01 -1.43164909e+00
-1.42025006e+00 8.94434452e-01 -2.44942918e-01 -6.06690466e-01
1.23536038e+00 -4.92578268e-01 -2.46052697e-01 -3.54569376e-01
-6.48012310e-02 2.20879078e-01 1.54331124e+00 -1.08503485e+00
-1.53353319e-01 -8.48806262e-01 -8.18874478e-01 -2.13746741e-01
9.12995562e-02 1.31216824e-01 -4.37833279e-01 -6.17972791e-01
4.26711470e-01 -1.00562382e+00 -3.60384583e-01 -2.19040290e-01
-6.37684882e-01 -1.57997757e-01 5.66033900e-01 -6.33738995e-01
9.52806532e-01 -2.20206571e+00 1.05550261e-02 8.88196111e-01
1.56439587e-01 1.03447609e-01 6.87040016e-02 -7.40917772e-02
-2.90017933e-01 -7.75407106e-02 -3.96940678e-01 -6.88979685e-01
1.50756329e-01 1.39414549e-01 -1.14316382e-01 1.16172218e+00
2.34430600e-02 4.09282267e-01 -5.70304811e-01 -1.82252750e-01
5.78379929e-02 6.33859158e-01 -3.63814920e-01 1.61702067e-01
-7.11098686e-02 6.63475871e-01 -4.56495248e-02 5.82373321e-01
7.90127397e-01 5.10954857e-01 -2.88779616e-01 1.13975830e-01
1.41788021e-01 -3.52013528e-01 -1.47622716e+00 1.11010969e+00
-2.20105484e-01 5.62453806e-01 5.64161651e-02 -6.43215358e-01
1.25804925e+00 1.81140974e-01 5.02364933e-01 -3.64405334e-01
5.93659937e-01 5.30878723e-01 -2.05808923e-01 -2.58028805e-01
2.57251859e-01 1.75424233e-01 7.20434710e-02 6.55231997e-02
6.45317435e-02 1.11797296e-01 7.02861696e-02 -3.42863530e-01
8.87960732e-01 6.50091544e-02 -9.14060995e-02 3.86854447e-02
5.83668232e-01 -8.27871501e-01 6.83374405e-01 7.22850502e-01
-2.08533257e-01 1.02192903e+00 6.24213815e-01 -3.92366588e-01
-7.53630579e-01 -9.48079228e-01 -1.88136131e-01 1.04190230e+00
-4.09533978e-01 -6.88694045e-02 -9.25877512e-01 -8.62501383e-01
1.12290949e-01 5.44528186e-01 -8.33640099e-01 -6.17578626e-01
-6.71916783e-01 -1.14784455e+00 4.93781030e-01 3.23313504e-01
8.71239826e-02 -1.10785258e+00 -3.84846330e-01 1.94358319e-01
2.63507515e-01 -8.60387981e-01 -3.83459598e-01 2.94882745e-01
-8.32754731e-01 -1.13115811e+00 -7.58877814e-01 -5.55955470e-01
1.11383843e+00 -6.11094415e-01 9.38975334e-01 1.84763819e-01
-2.99080282e-01 2.83249915e-01 -1.51806086e-01 -4.05929118e-01
-5.07499754e-01 -5.15997373e-02 2.71869838e-01 5.27918100e-01
5.50564706e-01 -5.63862383e-01 -2.96070814e-01 3.60059232e-01
-6.95175231e-01 -9.56872225e-01 2.31122702e-01 5.92532754e-01
5.06124318e-01 -1.35714244e-02 3.69696081e-01 -8.86348426e-01
7.29301155e-01 -4.91042703e-01 -7.10155010e-01 -1.37246624e-01
-4.21434164e-01 2.22419143e-01 5.48456311e-01 -5.65728664e-01
-4.83380407e-01 3.77866477e-01 -5.19364119e-01 -4.98248547e-01
-4.68544424e-01 1.05504967e-01 -3.53230506e-01 -2.27794439e-01
8.00205529e-01 1.04938395e-01 -9.44653805e-03 -5.49992502e-01
2.48264045e-01 4.29145157e-01 5.90073764e-01 -3.84361178e-01
8.24808061e-01 4.97383088e-01 7.67127723e-02 -1.03524458e+00
-5.34423470e-01 -4.31530684e-01 -7.95216501e-01 -3.41431767e-01
5.57589471e-01 -5.49838185e-01 -5.69647491e-01 9.29733574e-01
-8.43924701e-01 -3.42154086e-01 -3.18825394e-01 3.06061596e-01
-5.32571614e-01 4.44160044e-01 -2.81349272e-01 -8.96925926e-01
-3.57085437e-01 -1.00722206e+00 8.85482192e-01 1.46559015e-01
-3.97629887e-01 -8.14230204e-01 1.17391422e-02 -1.46025062e-01
2.50569016e-01 8.37698579e-01 3.66793364e-01 -9.73341763e-01
4.36638109e-02 -7.55158305e-01 1.75956622e-01 6.32240653e-01
-1.69127062e-01 5.29277980e-01 -1.10043120e+00 -4.75486666e-01
5.72257116e-02 -8.95395651e-02 9.71277595e-01 6.18751705e-01
1.18385184e+00 -2.10337967e-01 -6.23095445e-02 8.73920739e-01
1.14264250e+00 -1.67814448e-01 6.12187624e-01 3.22371334e-01
5.98164082e-01 6.72645867e-01 3.42769235e-01 5.60322106e-01
-3.97947282e-02 5.76953650e-01 6.80998087e-01 2.27997769e-02
2.63415247e-01 -6.59256503e-02 4.83322293e-01 2.80742168e-01
-6.29262999e-02 6.30652010e-02 -7.01027870e-01 4.30250734e-01
-1.85539722e+00 -6.32228851e-01 -2.32197210e-01 2.62222099e+00
4.60419744e-01 3.37805063e-01 4.33634341e-01 3.34807724e-01
7.52642274e-01 -1.48255061e-02 -5.82679272e-01 -4.32105809e-01
-1.10845968e-01 1.64085791e-01 6.84973478e-01 4.55088913e-01
-1.11896455e+00 9.13673043e-01 5.75572538e+00 4.75385487e-01
-1.29668760e+00 -4.34114374e-02 6.10405624e-01 -1.25094533e-01
1.92434430e-01 -6.23413146e-01 -8.26688468e-01 7.33802617e-01
8.88916373e-01 3.99479330e-01 7.26212487e-02 9.42513824e-01
2.77096748e-01 -1.17745936e-01 -9.68294203e-01 9.71207023e-01
4.22272921e-01 -4.85480517e-01 -3.82963270e-01 -3.25614214e-02
5.15670419e-01 1.09190568e-01 3.38913411e-01 2.35745564e-01
2.77736515e-01 -1.02196026e+00 8.23868871e-01 5.83406329e-01
2.75245547e-01 -9.42115545e-01 8.73971641e-01 2.36835063e-01
-6.61744952e-01 -4.48913798e-02 -4.38138455e-01 5.59457302e-01
-1.47387058e-01 7.95250416e-01 -8.10704052e-01 8.89262185e-02
7.30281532e-01 4.07518774e-01 -7.02869236e-01 1.35446835e+00
-3.84816855e-01 4.08861160e-01 -7.12904453e-01 1.31931439e-01
-2.09631454e-02 -2.06783056e-01 6.91857100e-01 1.12056124e+00
3.65387917e-01 -6.53624833e-01 -2.62181573e-02 6.05020940e-01
3.37912105e-02 4.29125577e-01 -4.87558544e-01 7.26768970e-01
1.43399462e-01 9.35301900e-01 -6.66491389e-01 1.80633694e-01
-1.02857403e-01 1.10919178e+00 2.75043011e-01 4.65686500e-01
-8.03130448e-01 -1.49815693e-01 4.95274425e-01 2.32951477e-01
3.50504994e-01 -8.73296857e-02 -2.13577911e-01 -9.36642051e-01
3.46900076e-01 -7.45867252e-01 3.56358498e-01 -3.77074331e-01
-1.30617845e+00 9.15376484e-01 -2.89669782e-01 -9.15690541e-01
-4.15697366e-01 -5.09778380e-01 -8.16679478e-01 6.42566741e-01
-1.33958900e+00 -7.69629896e-01 -2.68560201e-01 7.74701476e-01
2.04526365e-01 -2.24380642e-01 3.92275095e-01 2.58026958e-01
-8.09830070e-01 9.54553843e-01 8.01994279e-02 2.18406394e-01
8.34474087e-01 -1.35654998e+00 2.07416505e-01 9.89941418e-01
8.29569101e-02 4.42955375e-01 1.07634199e+00 -5.60726285e-01
-6.95457697e-01 -1.29380512e+00 7.66639471e-01 -3.44147623e-01
2.39272013e-01 -3.57844591e-01 -1.08609354e+00 5.93314886e-01
-4.98881638e-01 3.11086625e-01 5.63723803e-01 1.27440974e-01
-3.71996373e-01 -2.66688257e-01 -1.51160920e+00 5.60768843e-01
6.14073038e-01 -1.08606771e-01 -3.24397266e-01 2.84560829e-01
5.29111736e-02 -4.92882818e-01 -6.22492671e-01 4.65838373e-01
4.79780823e-01 -1.16930318e+00 7.28085935e-01 -4.57780600e-01
-3.63747299e-01 -2.13529348e-01 4.57521528e-02 -1.36774600e+00
1.32792354e-01 -7.60219574e-01 -3.32865387e-01 1.16537106e+00
3.15792888e-01 -6.10960782e-01 9.28910732e-01 6.60800338e-01
1.95801426e-02 -5.67707896e-01 -1.26436579e+00 -6.81904435e-01
-7.06094354e-02 -5.73496819e-01 4.31559920e-01 6.62446022e-01
-5.05138457e-01 -1.73483327e-01 -4.26656306e-01 4.05624747e-01
7.72846639e-01 -5.35280824e-01 1.01300240e+00 -1.47046947e+00
2.02512462e-02 -4.75782186e-01 -7.38060653e-01 -3.96589518e-01
4.46604252e-01 -4.33968335e-01 4.39875603e-01 -9.27646399e-01
-5.60109913e-01 -2.64028162e-01 -2.27697119e-01 2.63555497e-01
-1.22372225e-01 4.37183768e-01 -5.77701181e-02 -7.93625340e-02
-3.96028221e-01 5.30640841e-01 4.44506526e-01 1.85501575e-01
-4.83555049e-01 7.61682749e-01 -3.25580329e-01 1.25690985e+00
7.89781868e-01 -7.47284651e-01 1.83114693e-01 -1.07212905e-02
2.16013968e-01 -4.87112910e-01 3.21204573e-01 -1.20029593e+00
1.04382440e-01 3.55235964e-01 7.07367003e-01 -4.57840353e-01
2.42448464e-01 -1.06503332e+00 -1.36429980e-01 3.14423800e-01
-4.85065468e-02 8.26254934e-02 5.91942519e-02 4.68328238e-01
-3.08703650e-02 -3.75730008e-01 1.12933660e+00 6.37549534e-02
-3.76095682e-01 3.72540146e-01 -4.74239439e-01 -4.35073748e-02
9.14078176e-01 -2.98453033e-01 2.66129583e-01 -4.04567510e-01
-1.12531495e+00 1.39186746e-02 4.35238630e-01 4.41898435e-01
6.51231170e-01 -1.07613742e+00 -8.32755864e-01 5.81823647e-01
-3.07050515e-02 -2.22669952e-02 -3.38983424e-02 9.48665261e-01
-3.57642233e-01 -3.71860683e-01 1.13916107e-01 -6.70145988e-01
-1.49521601e+00 4.34033573e-01 6.61433101e-01 1.45476460e-01
-4.27175105e-01 8.86493862e-01 -4.49871242e-01 -5.18160164e-01
8.71410012e-01 -4.33644712e-01 -2.16142029e-01 2.81911820e-01
4.46354389e-01 5.12127399e-01 3.79160225e-01 -9.20955241e-01
-3.86481434e-01 6.21806264e-01 -5.99030172e-04 9.08426419e-02
1.26959586e+00 3.94950993e-02 -1.59535095e-01 4.89168376e-01
1.21820533e+00 2.89862663e-01 -1.28247654e+00 -1.57573298e-02
4.51098204e-01 -3.88293147e-01 -7.53816813e-02 -4.48758245e-01
-1.29318821e+00 4.57307070e-01 8.60684216e-01 2.80138757e-02
9.96872663e-01 -1.52983904e-01 5.50788522e-01 1.34488955e-01
-6.54559061e-02 -1.28021824e+00 -7.13014677e-02 4.13296700e-01
8.32734704e-01 -1.22118425e+00 1.19338473e-02 -4.66089919e-02
-3.81535351e-01 1.21769023e+00 6.50082231e-01 -4.78754044e-01
7.22434223e-01 1.99975327e-01 3.49495292e-01 -7.12496489e-02
-2.51107812e-01 -2.48117074e-01 4.99197274e-01 5.83051622e-01
2.09115505e-01 -2.32604697e-01 -2.00208202e-01 2.53365189e-01
-2.02833265e-01 -2.33566478e-01 3.68996024e-01 6.41226411e-01
-3.74719560e-01 -8.19060206e-01 -7.17217147e-01 2.89437532e-01
-6.11086547e-01 2.93746352e-01 -4.79950547e-01 7.10633755e-01
3.78627241e-01 7.51972914e-01 1.28750131e-01 -3.28665860e-02
5.75248539e-01 3.42930943e-01 4.38481957e-01 -3.13152492e-01
-5.46404958e-01 2.45786294e-01 -3.15086961e-01 -6.38216138e-01
-1.44676521e-01 -9.12543297e-01 -9.63261008e-01 -1.59936115e-01
-4.17270809e-01 3.30682918e-02 8.79147172e-01 9.31522846e-01
-1.02218375e-01 2.68891454e-01 5.79768479e-01 -8.38836014e-01
-7.25335002e-01 -9.74052906e-01 -6.02053702e-01 5.35286665e-01
6.67572916e-01 -5.70659876e-01 -7.40214109e-01 -2.99507588e-01]
|
[13.263623237609863, 0.496972918510437]
|
53bcd4ee-907b-4bb9-bdc8-eddf43331e9f
|
multi-modal-wireless-flexible-gel-free
|
2305.17629
| null |
https://arxiv.org/abs/2305.17629v1
|
https://arxiv.org/pdf/2305.17629v1.pdf
|
Multi-Modal Wireless Flexible Gel-Free Sensors with Edge Deep Learning for Detecting and Alerting Freezing of Gait in Parkinson's Patients
|
Freezing of gait (FoG) is a debilitating symptom of Parkinson's disease (PD). This work develops flexible wearable sensors that can detect FoG and alert patients and companions to help prevent falls. FoG is detected on the sensors using a deep learning (DL) model with multi-modal sensory inputs collected from distributed wireless sensors. Two types of wireless sensors are developed, including: (1) a C-shape central node placed around the patient's ears, which collects electroencephalogram (EEG), detects FoG using an on-device DL model, and generates auditory alerts when FoG is detected; (2) a stretchable patch-type sensor attached to the patient's legs, which collects electromyography (EMG) and movement information from accelerometers. The patch-type sensors wirelessly send collected data to the central node through low-power ultra-wideband (UWB) transceivers. All sensors are fabricated on flexible printed circuit boards. Adhesive gel-free acetylene carbon black and polydimethylsiloxane electrodes are fabricated on the flexible substrate to allow conformal wear over the long term. Custom integrated circuits (IC) are developed in 180 nm CMOS technology and used in both types of sensors for signal acquisition, digitization, and wireless communication. A novel lightweight DL model is trained using multi-modal sensory data. The inference of the DL model is performed on a low-power microcontroller in the central node. The DL model achieves a high detection sensitivity of 0.81 and a specificity of 0.88. The developed wearable sensors are ready for clinical experiments and hold great promise in improving the quality of life of patients with PD. The proposed design methodologies can be used in wearable medical devices for the monitoring and treatment of a wide range of neurodegenerative diseases.
|
['Xilin Liu', 'Thomas Dell', 'Yi Zhu', 'Jack Ji', 'Yuhan Hou']
|
2023-05-28
| null | null | null | null |
['electromyography-emg', 'eeg', 'specificity', 'eeg']
|
['medical', 'methodology', 'natural-language-processing', 'time-series']
|
[ 2.62442559e-01 -7.23012760e-02 -8.14332366e-02 -2.74073668e-02
-2.03917146e-01 -4.90095951e-02 -6.74964488e-01 -1.49958357e-01
-4.72385973e-01 8.83174598e-01 2.31694654e-01 3.08482260e-01
1.41714156e-01 -7.77332544e-01 -4.03519392e-01 -7.97116399e-01
-4.57788438e-01 -8.32895711e-02 4.57929164e-01 -1.92520134e-02
-4.71292138e-01 7.75724975e-03 -1.38893175e+00 2.36619756e-01
9.61377919e-01 1.42882276e+00 7.10490465e-01 2.45868430e-01
8.03159714e-01 2.96578079e-01 -8.96984935e-01 4.12714124e-01
-3.49330813e-01 6.04757778e-02 2.20731303e-01 -3.93611014e-01
-4.32603776e-01 -8.37226629e-01 -7.28231892e-02 8.74752581e-01
8.91229689e-01 -6.93051219e-01 4.35016155e-01 -1.04158580e+00
-4.19369638e-01 2.72204936e-01 -2.56531298e-01 5.05915880e-02
5.86619735e-01 3.53450656e-01 2.75571793e-01 -4.49590176e-01
3.22632909e-01 8.32180560e-01 1.04763961e+00 9.69892204e-01
-6.80856705e-01 -5.51295638e-01 -2.71085232e-01 3.12681705e-01
-8.41052473e-01 4.70016450e-02 7.25183010e-01 -3.33560616e-01
1.36807668e+00 1.70363054e-01 1.53718019e+00 1.54060221e+00
1.47746909e+00 2.82922775e-01 9.74862874e-01 6.96386695e-02
8.41287136e-01 -1.70581013e-01 3.93628567e-01 4.03212905e-01
1.06055069e+00 -1.32609621e-01 -5.21181345e-01 -3.57486069e-01
6.73767805e-01 6.09826326e-01 -5.42186797e-01 5.07476568e-01
-8.38622510e-01 3.85684669e-01 7.37199306e-01 3.99458498e-01
-7.38159359e-01 3.31625164e-01 2.97758371e-01 -2.57310569e-02
2.19196707e-01 -6.28279448e-02 -3.27627510e-01 -3.57465923e-01
-3.61735642e-01 -4.51638289e-02 8.18994164e-01 7.00541019e-01
-3.16463083e-01 3.32908481e-02 -9.63298306e-02 5.46879113e-01
1.02264440e+00 1.15819216e+00 8.93203676e-01 -5.13618052e-01
3.47260743e-01 7.18940437e-01 3.55309784e-01 -3.99929345e-01
-1.17091262e+00 -1.48449531e-02 -6.53857768e-01 2.28125602e-01
-3.44669193e-01 -6.99146986e-01 -5.93423367e-01 1.40376937e+00
1.50887012e-01 -2.49114752e-01 4.01976407e-02 1.12285566e+00
1.04256880e+00 4.23421353e-01 1.67330220e-01 -3.03502411e-01
1.61108160e+00 -5.39402338e-03 -1.00742197e+00 -5.94181538e-01
2.58352697e-01 5.53797297e-02 1.24568665e+00 5.88085055e-01
-6.86906040e-01 -2.53724098e-01 -1.61942005e+00 1.15190893e-01
-1.32069141e-01 2.38100305e-01 1.56140789e-01 9.10314143e-01
-8.91578913e-01 5.53799987e-01 -1.46467197e+00 -5.12065232e-01
3.58900517e-01 8.25192153e-01 1.80286601e-01 1.59107462e-01
-1.23498118e+00 9.62463140e-01 -3.17700714e-01 2.60831475e-01
-4.12914246e-01 -3.40294003e-01 -4.06023562e-01 -4.72869068e-01
-9.85235989e-01 -1.34888875e+00 8.00596356e-01 -2.02439979e-01
-1.83255279e+00 5.71957767e-01 1.06357224e-01 -5.74066222e-01
2.89945632e-01 -6.43785179e-01 -8.72514188e-01 3.65737736e-01
4.82689887e-02 2.64351666e-01 1.01543963e+00 -4.56869006e-01
-1.74585700e-01 -1.12329698e+00 -5.20959258e-01 -2.75819600e-01
-8.20402443e-01 -2.08052307e-01 7.36744761e-01 -5.06380200e-01
1.58390611e-01 -9.45967615e-01 3.04263800e-01 2.98053086e-01
-3.11489105e-01 -1.54256344e-01 1.45707965e+00 -6.46525383e-01
1.13101459e+00 -1.95416343e+00 -9.48497057e-02 8.86205882e-02
3.99011046e-01 3.49149972e-01 5.01179338e-01 1.94792673e-01
4.72590059e-01 -5.61060131e-01 -2.27985382e-01 -6.82206452e-02
-1.31768495e-01 4.48600590e-01 2.86238074e-01 9.08275127e-01
-1.08035363e-01 6.91789985e-01 -5.66793025e-01 2.32432246e-01
3.25524539e-01 8.34453285e-01 -1.12690613e-01 9.95558500e-02
2.82792305e-03 3.96653295e-01 -5.44446051e-01 1.07577682e+00
2.75123686e-01 -1.44126145e-02 -3.81113663e-02 -3.57541919e-01
1.51518434e-01 5.10859430e-01 -8.78369629e-01 1.46887958e+00
-9.65851545e-02 -5.75947277e-02 2.68907517e-01 -5.06736994e-01
1.15147424e+00 5.45030355e-01 6.40698373e-01 -8.99166644e-01
5.10046303e-01 5.20521700e-01 -2.04429328e-01 -1.35800505e+00
-3.59872878e-01 -2.89397806e-01 -1.46204144e-01 2.09103182e-01
-1.24054343e-01 3.16649914e-01 -4.24097717e-01 -4.73332107e-01
1.73152113e+00 -8.92409980e-02 -6.64415434e-02 -4.35968965e-01
-5.57031631e-02 -2.67629147e-01 8.85413408e-01 2.06404421e-02
-4.81941134e-01 3.88497412e-02 -1.63769707e-01 -3.77906710e-01
-4.76766407e-01 -1.35013556e+00 -1.22940332e-01 4.20541912e-01
3.61171216e-01 -3.16583991e-01 -7.67396212e-01 1.69619575e-01
4.36705738e-01 2.02579588e-01 -1.75039887e-01 -6.53337419e-01
-3.06819469e-01 -9.26977754e-01 4.34867680e-01 9.65636075e-01
6.88686907e-01 -8.67036104e-01 -1.61846447e+00 6.64692640e-01
-8.24766830e-02 -1.12876630e+00 1.58324525e-01 3.70803714e-01
-1.33241236e+00 -8.69389594e-01 -5.87683499e-01 -7.61927903e-01
2.13332489e-01 -2.79183596e-01 3.04886132e-01 -6.00089490e-01
-3.12308848e-01 7.44952619e-01 -3.59134942e-01 -1.03239489e+00
2.41386116e-01 -4.20977294e-01 8.35616708e-01 -3.02515835e-01
7.81531274e-01 -1.19084942e+00 -1.17434680e+00 2.43704125e-01
-2.51401037e-01 -5.33741653e-01 4.12841529e-01 2.65279800e-01
5.55492401e-01 -4.39130634e-01 9.67549205e-01 -2.58768946e-02
9.85478103e-01 -4.82358962e-01 -1.75476577e-02 -2.71760374e-01
-1.44868135e-01 -3.77429247e-01 5.67245245e-01 -5.71639538e-01
-3.75630289e-01 1.68681554e-02 -3.09565663e-01 8.99006948e-02
4.57007736e-02 2.28906274e-01 -4.78579402e-01 -1.13283262e-01
8.53338242e-01 -3.10107768e-01 2.39215761e-01 -2.73993343e-01
-3.68927091e-01 1.38324177e+00 7.50231206e-01 -1.07330956e-01
-8.32570195e-02 4.04008031e-01 -1.78603470e-01 -1.12891567e+00
2.04030886e-01 -1.88136116e-01 6.55315146e-02 -6.80048525e-01
1.07784855e+00 -1.45295250e+00 -8.54378879e-01 7.84614325e-01
-1.03190887e+00 -5.06948590e-01 -2.95671984e-03 9.17356193e-01
-4.66525674e-01 -1.38528243e-01 -1.00672746e+00 -7.79731154e-01
-1.36824453e+00 -7.80897498e-01 9.63377535e-01 2.65321940e-01
-1.06965792e+00 -5.94621360e-01 1.13703169e-01 2.57416576e-01
4.74964947e-01 8.56575608e-01 6.92293823e-01 3.54549974e-01
5.78841083e-02 -5.62537968e-01 5.58314145e-01 2.45519444e-01
4.87409830e-01 -5.42793989e-01 -1.08252990e+00 -4.80786830e-01
8.51574004e-01 -1.56245708e-01 5.14882624e-01 9.75157559e-01
4.65610594e-01 -9.06896293e-02 -8.25942397e-01 3.39456409e-01
1.40586627e+00 5.05380332e-01 1.00234985e+00 2.20416829e-01
6.23222589e-01 -2.89911985e-01 8.10875446e-02 5.78551233e-01
3.77278328e-01 1.93651110e-01 7.24373817e-01 3.14334422e-01
9.98318940e-02 1.85456991e-01 1.18401682e+00 9.40690279e-01
-3.75450104e-01 -1.73941940e-01 -3.78808945e-01 2.05510646e-01
-1.53220856e+00 -4.74432141e-01 -5.88282406e-01 2.10430884e+00
4.85495329e-01 2.35287264e-01 5.22644818e-01 5.62192082e-01
6.64061725e-01 -6.02596402e-01 -1.25831223e+00 -3.85247171e-01
9.38611478e-02 5.83315790e-01 5.78700960e-01 -6.24692775e-02
-7.93391883e-01 -2.72044893e-02 5.56608629e+00 -3.26013803e-01
-1.35548437e+00 5.39041817e-01 -2.58215994e-01 -6.64278388e-01
4.07084152e-02 -9.21026766e-01 -6.27690196e-01 1.18335676e+00
1.19187057e+00 3.81555796e-01 -1.12446301e-01 7.09148347e-01
6.43627524e-01 -3.87404114e-01 -1.02295673e+00 1.11893332e+00
-2.87317187e-01 -5.92170119e-01 -5.11484206e-01 -9.79871675e-02
9.54650750e-04 4.61273611e-01 -1.37249082e-01 -4.60204661e-01
-4.79844779e-01 -5.10567427e-01 5.95234811e-01 8.09085488e-01
1.04940057e+00 -5.10021448e-01 7.16397285e-01 1.68595865e-01
-1.13184369e+00 -6.76201224e-01 -4.32933211e-01 -5.86807013e-01
2.44551152e-01 1.35703897e+00 -2.35440925e-01 -1.96061879e-01
1.31752574e+00 8.41387570e-01 9.21160802e-02 8.74017417e-01
-5.05408049e-01 6.91678345e-01 -9.71710443e-01 -6.21805727e-01
-2.92295396e-01 -2.84228753e-02 7.83901930e-01 7.23105967e-01
5.96279442e-01 2.04413667e-01 -3.46763045e-01 1.01607370e+00
2.39620209e-01 -5.77250421e-01 -6.34970665e-01 2.76466638e-01
3.97708088e-01 8.93965662e-01 -3.63347739e-01 -1.01285521e-02
-3.61744791e-01 8.46985817e-01 -4.38205451e-01 3.12497783e-02
-7.08351016e-01 -5.41003108e-01 6.81794763e-01 5.05472660e-01
-2.64043082e-02 -3.92832816e-01 -6.68041527e-01 -8.70615602e-01
8.49021196e-01 -1.52548075e-01 4.87790816e-02 -8.18971276e-01
-1.43181610e+00 2.22809047e-01 -6.25585914e-01 -1.56713748e+00
2.00311422e-01 -7.08650708e-01 -7.22034216e-01 5.89941323e-01
-7.82565713e-01 -5.17383635e-01 -6.35148644e-01 1.06624079e+00
-1.23215497e-01 2.47546375e-01 1.27128243e+00 2.29077324e-01
-6.07168794e-01 1.92610085e-01 3.11048985e-01 -3.79109412e-01
6.32637560e-01 -7.43312001e-01 -5.03761359e-02 4.22777593e-01
-1.00739241e+00 4.33302194e-01 5.10633826e-01 -1.15332997e+00
-1.89238322e+00 -1.29676938e+00 4.70527828e-01 4.30220962e-02
3.42523277e-01 -5.00021935e-01 -5.85797429e-01 3.22655439e-01
-2.29742199e-01 -9.50092375e-02 7.82274306e-01 -6.75003111e-01
3.65179300e-01 -6.40572548e-01 -2.06536245e+00 7.05177188e-02
1.03735816e+00 -2.66214937e-01 -8.75058651e-01 5.82705438e-01
3.37300748e-01 -3.54278982e-01 -1.31377423e+00 3.97076458e-01
1.08240139e+00 -4.40567613e-01 5.89239120e-01 2.87186921e-01
-6.88146502e-02 -2.67980635e-01 -1.84952378e-01 -1.35353756e+00
-1.81506023e-01 -4.73938644e-01 -5.24430275e-01 6.91054702e-01
-2.98780985e-02 -1.33028746e+00 4.80031461e-01 6.15124881e-01
-4.39350784e-01 -8.75054836e-01 -1.23570848e+00 -9.69267607e-01
-4.36103910e-01 -1.62308812e-01 2.00899258e-01 7.87935778e-02
1.13445139e+00 3.04695696e-01 1.99126415e-02 2.70961851e-01
5.39220750e-01 -5.91810822e-01 -2.41763890e-01 -1.64512897e+00
8.62130597e-02 5.11344671e-01 -1.12839484e+00 -5.33793271e-01
-5.92201352e-01 -4.90209728e-01 -2.64539290e-02 -2.00475955e+00
-4.40410227e-01 6.10644408e-02 -3.17629188e-01 4.43626583e-01
3.47072273e-01 1.32161111e-01 -3.52144420e-01 1.09592006e-01
-1.58272937e-01 4.48707461e-01 8.29586744e-01 -2.66459167e-01
-6.58757091e-01 1.78569004e-01 -2.87815094e-01 8.02870691e-01
1.16531551e+00 -4.78256017e-01 -4.11592841e-01 -5.85837603e-01
2.00466841e-01 -6.65696785e-02 6.87855422e-01 -1.88451946e+00
2.68853575e-01 7.49438703e-01 8.47084045e-01 -1.64325461e-01
4.37311113e-01 -1.05218422e+00 3.60543013e-01 1.16494298e+00
6.43531561e-01 9.03433859e-02 7.52298161e-02 5.22251606e-01
3.86674821e-01 4.73487735e-01 8.59165847e-01 -1.15116658e-02
-1.45979226e-01 -1.15395278e-01 -1.17221975e+00 -5.63431799e-01
1.07834792e+00 -6.06382966e-01 -5.10386050e-01 5.66429691e-03
-1.03273094e+00 7.75148869e-02 1.21507689e-01 1.42545223e-01
9.79286134e-01 -1.45332265e+00 1.10477053e-01 4.20565844e-01
-2.02018738e-01 -1.67394519e-01 2.39364952e-01 1.11140800e+00
-2.67260462e-01 2.30460986e-01 -6.73397303e-01 -1.01502848e+00
-1.01889277e+00 -1.61433846e-01 4.94027525e-01 5.53380489e-01
-1.21148288e+00 5.29751122e-01 -1.01734221e+00 4.96252865e-01
3.12167168e-01 -1.12170446e+00 -9.27818939e-02 -6.43677190e-02
8.29233646e-01 7.45084524e-01 4.33505923e-01 -5.38059399e-02
-8.51420999e-01 9.05891478e-01 4.53510672e-01 1.63670063e-01
1.74027586e+00 1.15299467e-02 -3.67373116e-02 7.65994489e-01
8.51554930e-01 -7.20010281e-01 -1.12570059e+00 3.91611934e-01
-4.25938129e-01 4.14339632e-01 -4.62512933e-02 -9.53924179e-01
-9.79667425e-01 5.31241417e-01 1.56261551e+00 5.41184470e-02
1.38059402e+00 -2.00547040e-01 1.46852458e+00 3.15618157e-01
1.03825033e+00 -1.36794603e+00 -1.23448960e-01 -1.28015861e-01
8.61313224e-01 -5.07485986e-01 -6.72215745e-02 -2.01342806e-01
-5.67959249e-02 1.19479668e+00 3.35478574e-01 -8.20124149e-01
8.74581158e-01 7.90625930e-01 -1.02328636e-01 -2.73299456e-01
-3.41091812e-01 3.07533085e-01 -4.48973209e-01 1.23398602e+00
1.92459002e-01 6.03079319e-01 -7.01294661e-01 1.56199658e+00
-2.58746266e-01 8.88230562e-01 3.34619880e-01 1.41016805e+00
-1.02743793e+00 -4.62904006e-01 -5.19695342e-01 1.06505930e+00
-3.01998943e-01 5.56484520e-01 -3.68144512e-01 1.81822076e-01
5.84461927e-01 1.26848257e+00 4.51809578e-02 -8.14054191e-01
8.04279089e-01 9.47487578e-02 7.57449448e-01 -2.75635600e-01
-4.81921941e-01 1.03251562e-01 -1.17872108e-03 -8.54237735e-01
-4.67804044e-01 -5.28891802e-01 -1.62586772e+00 1.56270772e-01
1.36023983e-01 -4.95739877e-01 7.37235606e-01 1.22944617e+00
6.66691065e-01 7.73151457e-01 2.31266946e-01 -9.25772190e-01
-1.82312548e-01 -1.27707064e+00 -1.15855849e+00 -2.47403815e-01
4.43311512e-01 -9.51664627e-01 -1.98525414e-01 -1.19797394e-01]
|
[13.569940567016602, 3.259120464324951]
|
3c5df030-b887-4926-adb5-02123825543b
|
sleep-model-a-sequence-model-for-predicting
|
2302.12709
| null |
https://arxiv.org/abs/2302.12709v1
|
https://arxiv.org/pdf/2302.12709v1.pdf
|
Sleep Model -- A Sequence Model for Predicting the Next Sleep Stage
|
As sleep disorders are becoming more prevalent there is an urgent need to classify sleep stages in a less disturbing way.In particular, sleep-stage classification using simple sensors, such as single-channel electroencephalography (EEG), electrooculography (EOG), electromyography (EMG), or electrocardiography (ECG) has gained substantial interest. In this study, we proposed a sleep model that predicts the next sleep stage and used it to improve sleep classification accuracy. The sleep models were built using sleep-sequence data and employed either statistical $n$-gram or deep neural network-based models. We developed beam-search decoding to combine the information from the sensor and the sleep models. Furthermore, we evaluated the performance of the $n$-gram and long short-term memory (LSTM) recurrent neural network (RNN)-based sleep models and demonstrated the improvement of sleep-stage classification using an EOG sensor. The developed sleep models significantly improved the accuracy of sleep-stage classification, particularly in the absence of an EEG sensor.
|
['Wonyong Sung', 'Iksoo Choi']
|
2023-02-17
| null | null | null | null |
['electromyography-emg', 'electrocardiography-ecg', 'eeg', 'eeg']
|
['medical', 'methodology', 'methodology', 'time-series']
|
[ 3.00883591e-01 -4.83576775e-01 2.45946366e-03 -6.63977504e-01
-4.59195614e-01 -7.18801692e-02 -4.54547778e-02 -5.31706512e-02
-7.81640291e-01 8.51297379e-01 2.02318460e-01 -5.21764040e-01
-9.58245695e-02 -3.93553078e-01 -1.24867246e-01 -6.23889685e-01
-1.12048782e-01 1.52745452e-02 -9.29982439e-02 -4.44009118e-02
3.69830787e-01 1.46325409e-01 -1.62277639e+00 1.38319805e-01
1.05707026e+00 1.45637023e+00 7.57879734e-01 7.93393552e-01
-3.04764155e-02 3.87491614e-01 -8.76755536e-01 2.39724994e-01
-1.64167926e-01 -7.41369069e-01 -5.09246111e-01 -3.71158361e-01
-3.27498108e-01 3.20702553e-01 4.90541384e-02 8.95494998e-01
7.10578859e-01 1.89935535e-01 1.20008290e-01 -7.68963039e-01
-2.74802744e-01 2.93756366e-01 1.12315472e-02 8.37843120e-01
5.57451606e-01 5.58190122e-02 6.18581176e-01 -5.08694768e-01
-2.88869571e-02 3.86441946e-01 8.28834832e-01 7.75226414e-01
-9.99664009e-01 -9.47459400e-01 -1.96931288e-01 7.40116894e-01
-1.42536569e+00 -5.31538367e-01 4.78764892e-01 -1.76060438e-01
1.68124032e+00 3.05566907e-01 1.31057084e+00 1.27395284e+00
1.31189048e+00 3.26248735e-01 1.41261148e+00 -3.89350682e-01
3.61799449e-01 6.50960505e-02 2.55646706e-01 4.93503451e-01
5.26867993e-02 1.57775208e-02 -8.25887263e-01 1.84979513e-01
1.68153763e-01 5.32717109e-01 -1.45027399e-01 6.11397088e-01
-8.56013358e-01 4.76300895e-01 3.05914938e-01 6.52793527e-01
-5.57970345e-01 -1.85898036e-01 3.14361364e-01 3.67013633e-01
6.44585788e-01 6.71915650e-01 -4.26927298e-01 -8.32269132e-01
-1.54743803e+00 -6.28195167e-01 8.34060967e-01 4.57506776e-01
4.73856986e-01 1.26463041e-01 1.44178554e-01 8.27031434e-01
2.99008071e-01 5.43510377e-01 1.30834651e+00 -4.67870802e-01
1.56512752e-01 7.67560184e-01 -2.28772640e-01 -6.24896467e-01
-1.01413572e+00 -3.66014510e-01 -1.11386740e+00 -2.41100401e-01
-2.32792109e-01 -4.04956602e-02 -7.86585629e-01 1.36939728e+00
-5.09092689e-01 3.22375268e-01 -6.92944601e-02 5.65154076e-01
6.38551712e-01 4.71025139e-01 6.00909665e-02 -5.43381214e-01
1.30953395e+00 -6.29905999e-01 -8.87183666e-01 -6.82476878e-01
4.48896319e-01 -3.62781137e-01 8.00844193e-01 5.66106081e-01
-1.01203930e+00 -6.62071586e-01 -1.16028273e+00 2.73423851e-01
-4.38486904e-01 1.79213397e-02 1.52928099e-01 7.23943114e-01
-1.38455260e+00 7.64540792e-01 -1.45034814e+00 -7.41527855e-01
-5.77150891e-03 9.40835834e-01 -3.09054464e-01 4.72443789e-01
-1.11886013e+00 1.11744714e+00 -9.06127915e-02 5.01281679e-01
-5.16754210e-01 7.72511885e-02 -8.57749879e-01 3.77537198e-02
-2.59369314e-01 -5.47945380e-01 7.92988837e-01 -6.42574489e-01
-1.64534879e+00 7.66670585e-01 -9.00284052e-01 -6.38882160e-01
-6.01612747e-01 -2.54938155e-01 -8.30780149e-01 -1.28912916e-02
-9.40649584e-02 3.08266431e-01 9.75140274e-01 -1.19719006e-01
-5.21678627e-01 -6.58257246e-01 -5.67360699e-01 -1.45638347e-01
-3.47653896e-01 1.87103555e-01 2.23710403e-01 -1.71049461e-01
1.36434764e-01 -1.13214576e+00 -2.00771078e-01 -6.98940933e-01
-3.11402082e-01 -2.37158522e-01 2.05957070e-01 -7.17666924e-01
1.81937099e+00 -2.26978087e+00 1.63727522e-01 9.70244929e-02
2.20950648e-01 3.33837777e-01 2.66051650e-01 2.81802505e-01
-9.79905501e-02 4.68930565e-02 -2.90315356e-02 -6.51959002e-01
-3.20573151e-01 5.36826313e-01 8.81055966e-02 4.13053185e-01
-2.27816075e-01 9.87027168e-01 -4.89720851e-01 -2.86158979e-01
3.35492939e-01 2.41489157e-01 -9.60154161e-02 4.49766636e-01
4.08888340e-01 4.43221897e-01 8.14275444e-02 4.88734722e-01
-1.16090491e-01 -4.81447667e-01 -6.15336634e-02 9.68347415e-02
-3.99440199e-01 8.44261646e-01 -3.63083541e-01 2.01474452e+00
-8.13453734e-01 8.21581960e-01 -2.57410437e-01 -7.88930178e-01
1.08544540e+00 3.34085137e-01 1.94888711e-01 -1.04637575e+00
2.90250570e-01 1.52700856e-01 3.25287998e-01 -6.35303795e-01
2.63034731e-01 -3.30550611e-01 -1.67858880e-02 4.93854910e-01
1.85526293e-02 2.34256849e-01 6.42502457e-02 -5.31967223e-01
1.28409195e+00 -2.44035050e-01 5.31298101e-01 -1.86882406e-01
4.09799457e-01 -5.03016651e-01 6.82449222e-01 5.97558200e-01
-6.42241538e-02 4.79129195e-01 -1.69691712e-01 -3.16408128e-01
-2.80784249e-01 -7.98152447e-01 -3.44987065e-02 9.03920352e-01
-2.15390846e-02 -8.87691379e-01 -5.68926871e-01 -1.30651325e-01
-3.37096840e-01 8.08711052e-01 -6.49925590e-01 -5.94320655e-01
-2.86295325e-01 -8.69595528e-01 5.19442260e-01 5.68874419e-01
3.17045778e-01 -1.27626038e+00 -1.09390211e+00 4.48636323e-01
-1.32229492e-01 -8.43090296e-01 -2.44237602e-01 1.10669696e+00
-1.17901540e+00 -7.70239413e-01 -5.30729890e-01 -5.87463021e-01
3.21324915e-01 -1.19492844e-01 7.82742143e-01 -1.88371390e-01
-2.52800494e-01 8.07479098e-02 -4.96469259e-01 -3.04115951e-01
-1.59585923e-02 4.16474432e-01 6.59732997e-01 -7.78583288e-02
9.51619446e-01 -1.10897601e+00 -8.63330722e-01 2.52306044e-01
-3.70940417e-01 -2.69552972e-02 7.61817157e-01 4.97844577e-01
5.05538523e-01 -1.48136899e-01 5.73609889e-01 -2.40867898e-01
9.19909298e-01 -4.73762274e-01 -1.72674656e-01 9.21882764e-02
-1.15931678e+00 8.53556767e-02 8.27309608e-01 -2.19419494e-01
-4.00255144e-01 -3.77298862e-01 -7.09836423e-01 -1.77010193e-01
-2.05580592e-01 5.46357870e-01 1.23162121e-01 5.02569303e-02
4.76541132e-01 8.32486153e-01 1.19975465e-03 -7.47116506e-01
-5.23274302e-01 1.30229545e+00 3.58090222e-01 1.65242985e-01
-8.02946556e-03 7.88752213e-02 -1.66162685e-01 -1.26847827e+00
-7.57647812e-01 -5.69168925e-01 -5.97597659e-01 -1.14528418e-01
1.22173798e+00 -8.07774246e-01 -9.08655345e-01 5.91037571e-01
-7.79394984e-01 -4.85400826e-01 -3.00938971e-02 8.33101273e-01
-2.85908908e-01 2.59819180e-01 -6.34634554e-01 -9.61215436e-01
-8.90990913e-01 -1.15952647e+00 8.22412789e-01 4.74215060e-01
-8.62393796e-01 -7.89386690e-01 1.92879453e-01 7.15543106e-02
6.15560293e-01 -2.62419969e-01 6.56731308e-01 -6.57478034e-01
-3.46168652e-02 -1.76592797e-01 2.96799928e-01 6.47878885e-01
4.20117587e-01 -6.22367144e-01 -7.71764517e-01 -1.88195199e-01
6.83255076e-01 7.34362304e-02 4.67451066e-01 5.54980040e-01
1.11616433e+00 -5.86568117e-02 -2.87328213e-01 9.87324238e-01
9.73937690e-01 7.50922918e-01 6.88415527e-01 4.34961766e-01
4.15717363e-01 -1.65413618e-01 9.40970629e-02 3.26569736e-01
4.31090087e-01 4.52117443e-01 -2.59623714e-02 2.13456869e-01
2.26017445e-01 1.84849817e-02 6.35846317e-01 1.54602933e+00
-1.53554395e-01 -7.31297731e-02 -8.44995499e-01 2.53596395e-01
-1.39587259e+00 -6.97973251e-01 1.08369611e-01 2.00438404e+00
7.22747684e-01 4.71790433e-01 7.41870552e-02 2.66937286e-01
2.35709801e-01 4.59221005e-02 -6.89027786e-01 -9.53013003e-01
1.51353583e-01 8.93153846e-01 4.44970518e-01 -4.47499715e-02
-4.70158368e-01 5.05398989e-01 6.83695507e+00 4.10990983e-01
-1.55751622e+00 3.46087754e-01 5.15410416e-02 -5.71506798e-01
2.80536622e-01 -4.13575292e-01 -9.37962055e-01 1.21170270e+00
1.97473967e+00 1.08254120e-01 7.05407917e-01 7.01577783e-01
6.29053116e-01 -6.84867740e-01 -1.03245521e+00 1.50177443e+00
2.51881957e-01 -9.16778922e-01 -5.51233649e-01 1.82273522e-01
7.51070753e-02 5.49124479e-01 -2.60092944e-01 2.59543329e-01
-5.26491821e-01 -1.13598919e+00 2.70537257e-01 9.69504535e-01
9.68932748e-01 -4.61279184e-01 8.94796550e-01 5.04562259e-01
-1.20081711e+00 -3.06948721e-01 -1.95026726e-01 -5.42286217e-01
2.17225149e-01 4.04306024e-01 -5.80599725e-01 1.78812176e-01
1.00672340e+00 1.16841471e+00 -8.68456781e-01 1.02710950e+00
-3.28621060e-01 9.72647905e-01 -4.38841283e-01 -4.24239904e-01
3.92975397e-02 -3.05799574e-01 2.07090318e-01 1.06301332e+00
5.91618478e-01 8.35850313e-02 -3.42503428e-01 7.15742886e-01
3.62586975e-01 -5.05622149e-01 -4.14416552e-01 -2.82341719e-01
9.61048827e-02 1.10573208e+00 -7.16322005e-01 5.96248284e-02
-5.10406017e-01 1.11757088e+00 -8.98755789e-02 1.82394326e-01
-3.12818140e-01 -6.68956935e-01 6.97519481e-01 -4.40640748e-02
3.92428786e-02 -2.84029603e-01 -4.91724759e-01 -1.21929538e+00
-1.19802475e-01 -4.97694492e-01 1.89819723e-01 -1.02341819e+00
-9.92666245e-01 9.77723360e-01 -2.05595627e-01 -9.83738720e-01
-5.78960836e-01 -3.93985569e-01 -9.00522828e-01 1.10148621e+00
-1.10681093e+00 -3.76290441e-01 -1.63464442e-01 5.65713406e-01
7.07759559e-01 -9.91856232e-02 1.03394079e+00 4.46684092e-01
-8.58404934e-01 4.99825239e-01 4.67775427e-02 -1.43053129e-01
4.74191755e-01 -1.25050402e+00 3.75750065e-01 7.36777604e-01
-1.72202978e-02 1.06887496e+00 5.50206780e-01 -5.69896519e-01
-1.34177125e+00 -7.61863947e-01 1.32213068e+00 -2.35193491e-01
4.70948189e-01 -4.62391853e-01 -5.74805021e-01 5.25693953e-01
1.74351841e-01 -6.52470171e-01 1.29167414e+00 1.99422583e-01
4.30981725e-01 -4.91741180e-01 -7.77979612e-01 1.90484419e-01
8.61407399e-01 -8.78124774e-01 -1.16078579e+00 -3.30358207e-01
1.92739323e-01 -8.32354501e-02 -7.80322790e-01 3.42987478e-01
8.12234163e-01 -1.07902408e+00 4.66158867e-01 -7.52698183e-02
6.57979399e-02 1.50213391e-01 -5.72729558e-02 -1.43288803e+00
-1.81019649e-01 -6.20124161e-01 -1.27671778e-01 5.62717140e-01
3.97185117e-01 -8.21258843e-01 7.00357616e-01 6.16851807e-01
-4.74026054e-01 -9.99057353e-01 -9.69270945e-01 -6.18673325e-01
-6.70845449e-01 -5.46861351e-01 3.86998117e-01 9.20951664e-02
5.60317338e-01 6.72527254e-01 -4.63277578e-01 -2.97455192e-01
1.22290999e-02 4.43103909e-02 1.52903482e-01 -1.13445055e+00
-8.56380314e-02 -2.18623772e-01 -6.32374644e-01 -1.08937705e+00
2.48701513e-01 -1.03166652e+00 1.99687079e-01 -1.85874188e+00
3.19155352e-03 4.59419861e-02 -1.01762521e+00 4.64086622e-01
-3.11801247e-02 4.10954833e-01 -6.37747720e-02 4.31384407e-02
-8.08283448e-01 4.50088143e-01 5.41761339e-01 2.29631245e-01
-5.08348107e-01 4.63835031e-01 -5.90739071e-01 6.41365707e-01
9.33854640e-01 -6.78814471e-01 -2.79708922e-01 -1.76419541e-01
5.81324220e-01 3.07385802e-01 -1.06106520e-01 -1.49275339e+00
6.53784335e-01 5.44986486e-01 6.50525510e-01 -6.35466099e-01
7.24371433e-01 -5.72147667e-01 1.61309898e-01 6.83258355e-01
-1.47258237e-01 4.69442308e-01 2.37428755e-01 3.39594752e-01
-6.95081949e-02 -1.64929613e-01 3.90195251e-01 -2.57668704e-01
-5.57870626e-01 -4.02136482e-02 -1.18664849e+00 -3.69065821e-01
4.59420562e-01 -7.96623468e-01 2.34604686e-01 -2.86005139e-01
-1.00702345e+00 1.24587566e-01 2.27356747e-01 3.58724087e-01
8.51754606e-01 -7.51613438e-01 -1.86904371e-02 8.46151173e-01
-5.36084622e-02 -4.67935055e-01 9.06411856e-02 1.34508276e+00
-9.86789390e-02 7.32720256e-01 -5.68256438e-01 -7.04689801e-01
-1.45634758e+00 2.10753366e-01 2.10493341e-01 -6.62538484e-02
-4.68174219e-01 9.03659284e-01 -5.02117991e-01 3.67463455e-02
3.03598732e-01 -8.99935365e-01 -4.27433938e-01 1.19741432e-01
6.37960136e-01 3.86658221e-01 3.69519711e-01 -4.15864140e-01
-7.35533357e-01 4.71166372e-01 2.04542264e-01 1.69078395e-01
1.41607547e+00 -4.97771293e-01 -3.85481089e-01 1.16219938e+00
1.15856147e+00 -4.14723724e-01 -4.77946848e-01 3.19749624e-01
2.05736160e-01 6.17686026e-02 1.90442771e-01 -7.96172500e-01
-6.44104302e-01 1.02309787e+00 9.58834648e-01 3.17164481e-01
1.47425818e+00 -2.74140358e-01 1.36355221e+00 5.24173439e-01
6.48065746e-01 -9.88774121e-01 -4.56173271e-01 4.36182350e-01
9.85573083e-02 -9.06732202e-01 -1.00464866e-01 6.17895126e-01
-3.23475868e-01 1.18465030e+00 2.06207693e-01 -7.02447072e-02
9.19688284e-01 1.34151697e-01 2.30739713e-02 -2.36944035e-01
-8.44616473e-01 -2.29427055e-01 3.31973046e-01 2.68394798e-01
4.38526690e-01 8.97238031e-02 -3.79108727e-01 9.66334283e-01
-5.90706170e-01 5.38113236e-01 2.45685473e-01 8.67677093e-01
-5.47529697e-01 -1.08654130e+00 -4.97173406e-02 1.25965869e+00
-7.33851433e-01 -3.80675852e-01 -1.68356031e-01 2.40752503e-01
1.12608284e-01 1.36834598e+00 1.64175227e-01 -1.07089353e+00
2.04919148e-02 5.69222927e-01 3.27271223e-01 -9.23433661e-01
-9.38662350e-01 -6.55413494e-02 -1.39747813e-01 -9.55880523e-01
-5.14927685e-01 -4.73565251e-01 -9.10530746e-01 -2.29749174e-04
-3.62361908e-01 5.55241168e-01 9.34686542e-01 1.32430649e+00
6.60617530e-01 6.32099032e-01 4.33642656e-01 -8.31354916e-01
2.07104301e-03 -1.35602403e+00 -9.29809928e-01 -2.61299342e-01
5.74237943e-01 -3.74332607e-01 -5.14247835e-01 -1.60256103e-01]
|
[13.536579132080078, 3.4910805225372314]
|
982dee6d-4a9d-4d36-a116-443deaa1a18d
|
weakly-supervised-segmentation-of-referring
|
2205.04725
| null |
https://arxiv.org/abs/2205.04725v2
|
https://arxiv.org/pdf/2205.04725v2.pdf
|
Weakly-supervised segmentation of referring expressions
|
Visual grounding localizes regions (boxes or segments) in the image corresponding to given referring expressions. In this work we address image segmentation from referring expressions, a problem that has so far only been addressed in a fully-supervised setting. A fully-supervised setup, however, requires pixel-wise supervision and is hard to scale given the expense of manual annotation. We therefore introduce a new task of weakly-supervised image segmentation from referring expressions and propose Text grounded semantic SEGgmentation (TSEG) that learns segmentation masks directly from image-level referring expressions without pixel-level annotations. Our transformer-based method computes patch-text similarities and guides the classification objective during training with a new multi-label patch assignment mechanism. The resulting visual grounding model segments image regions corresponding to given natural language expressions. Our approach TSEG demonstrates promising results for weakly-supervised referring expression segmentation on the challenging PhraseCut and RefCOCO datasets. TSEG also shows competitive performance when evaluated in a zero-shot setting for semantic segmentation on Pascal VOC.
|
['Cordelia Schmid', 'Ivan Laptev', 'Robin Strudel']
|
2022-05-10
| null | null | null | null |
['referring-expression-segmentation']
|
['computer-vision']
|
[ 8.22399139e-01 5.33142686e-01 -3.49217236e-01 -7.41687179e-01
-1.30130696e+00 -6.86833441e-01 3.91894132e-01 8.57993364e-02
-3.67534339e-01 4.60828722e-01 -2.40870103e-01 -1.13313191e-01
3.16145390e-01 -5.66493869e-01 -1.13467669e+00 -6.13038301e-01
6.92147553e-01 7.73633838e-01 5.65039456e-01 -2.03323826e-01
1.37068272e-01 2.63932377e-01 -1.36044550e+00 4.40683395e-01
8.41552019e-01 9.53494787e-01 2.82117784e-01 6.54716730e-01
-5.36586523e-01 8.57411742e-01 -4.19758052e-01 -4.96483833e-01
1.33495122e-01 -7.07186341e-01 -1.41839039e+00 7.62272596e-01
8.98976564e-01 2.95004249e-01 3.65487397e-01 1.26599944e+00
1.04962185e-01 4.02095094e-02 4.98042613e-01 -1.19216645e+00
-5.56435823e-01 5.20750463e-01 -6.91635549e-01 5.23585081e-03
2.23770037e-01 -7.26602748e-02 1.23365271e+00 -6.98642075e-01
9.95166063e-01 1.27128160e+00 6.77205503e-01 7.49304831e-01
-1.64911795e+00 -1.51335709e-02 5.04013956e-01 -2.04244673e-01
-1.23522162e+00 -8.01667273e-02 7.52911627e-01 -5.76280057e-01
7.88571715e-01 1.13973133e-01 3.86855870e-01 7.06452131e-01
-4.08622772e-01 1.23981154e+00 1.29874718e+00 -5.77203214e-01
3.34076077e-01 2.34763876e-01 2.88867265e-01 8.98147166e-01
-3.74410838e-01 -5.79249978e-01 -3.75556290e-01 2.02961296e-01
5.95099807e-01 -9.19740871e-02 -2.64577448e-01 -6.87527359e-01
-1.06466925e+00 8.20965827e-01 7.32425272e-01 4.99666721e-01
-2.50408918e-01 6.13519609e-01 3.18913460e-01 -2.10375637e-01
8.33120346e-01 2.61710107e-01 -5.44389784e-01 1.44096717e-01
-1.31027019e+00 7.10737407e-02 6.46828592e-01 1.31668997e+00
1.23501527e+00 -2.00950220e-01 -3.07843179e-01 8.27233911e-01
1.19656324e-02 3.92907619e-01 -5.49045727e-02 -1.05817807e+00
4.18286622e-01 8.09407175e-01 -4.94759232e-02 -6.42242551e-01
-3.63837719e-01 -2.93186814e-01 -2.88825184e-01 -8.16461667e-02
6.86391115e-01 -1.05135227e-02 -1.33869433e+00 1.61551297e+00
4.03224528e-01 -1.57369629e-01 -4.24160101e-02 1.10192502e+00
8.81050944e-01 6.71366334e-01 3.06561053e-01 1.79697365e-01
1.53179860e+00 -1.27550864e+00 -6.36825681e-01 -4.76613402e-01
8.23566735e-01 -4.89075214e-01 1.47845900e+00 1.89235941e-01
-1.18781269e+00 -3.29504609e-01 -6.50039673e-01 -4.13353562e-01
-5.06891012e-01 2.67682791e-01 5.73897719e-01 6.37208641e-01
-1.11938655e+00 2.11566672e-01 -8.25870931e-01 -7.37931490e-01
8.53620946e-01 2.98379064e-01 -2.80085623e-01 3.69854830e-02
-6.65302694e-01 5.53782582e-01 4.87922013e-01 3.45634595e-02
-8.64020705e-01 -6.08421981e-01 -1.04618764e+00 -2.68005226e-02
7.47115672e-01 -5.52915573e-01 1.18271077e+00 -1.57866228e+00
-1.21996522e+00 1.71779048e+00 -4.85659003e-01 -7.28246629e-01
4.50919658e-01 -1.28967240e-01 3.49632561e-01 6.24914289e-01
4.28348839e-01 1.42052925e+00 8.63559723e-01 -1.60091424e+00
-5.88142037e-01 -3.80152255e-01 1.87501535e-01 8.03927928e-02
2.67271519e-01 1.37779623e-01 -9.27808523e-01 -5.60125709e-01
3.42076987e-01 -8.39048445e-01 -3.96848321e-01 2.57320534e-02
-7.33153105e-01 -3.69169384e-01 9.96507585e-01 -5.28966665e-01
5.75564265e-01 -1.87914276e+00 3.78772438e-01 4.34717499e-02
8.43291208e-02 -5.00572920e-02 -2.72272639e-02 9.86931380e-04
-5.01827635e-02 2.89065152e-01 -8.83868933e-01 -6.52682006e-01
1.19273596e-01 5.52512169e-01 -2.49874011e-01 4.31640625e-01
4.96646285e-01 1.38024080e+00 -8.03355277e-01 -9.25839305e-01
2.96666652e-01 2.05518112e-01 -5.63567638e-01 2.56325722e-01
-9.69287276e-01 5.17926633e-01 -4.71626818e-01 9.19255197e-01
5.68332136e-01 -2.50485957e-01 -2.07098812e-01 -1.40730128e-01
5.22279516e-02 -3.14227462e-01 -6.01574004e-01 2.13580894e+00
-5.51726222e-01 7.27663159e-01 1.76605254e-01 -1.45991194e+00
9.29427743e-01 3.50491330e-02 3.93035680e-01 -5.81787705e-01
3.09154719e-01 2.06265152e-01 -6.56197011e-01 -5.21940231e-01
3.26369137e-01 -3.31766635e-01 -4.96376395e-01 2.82311082e-01
4.30740893e-01 -8.15055192e-01 3.08211267e-01 3.94500703e-01
6.50225639e-01 6.45145535e-01 6.96867406e-02 -4.99487877e-01
6.50740087e-01 3.78497720e-01 4.08571303e-01 6.62109673e-01
-1.91622183e-01 1.00145769e+00 9.58771348e-01 -3.25369090e-01
-8.53325784e-01 -8.43494654e-01 -2.45925441e-01 1.45868385e+00
3.38159025e-01 -1.17661037e-01 -1.44964612e+00 -1.06326091e+00
-3.59110236e-01 7.21145928e-01 -7.58041382e-01 3.68497044e-01
-4.30831164e-01 -3.86071146e-01 3.47017288e-01 5.88721693e-01
6.10420108e-01 -1.10708725e+00 -5.78480005e-01 -1.16279274e-01
-1.67945623e-01 -1.72740149e+00 -6.75271928e-01 5.91012120e-01
-7.16305733e-01 -8.77560318e-01 -1.04045498e+00 -1.22065663e+00
1.08918977e+00 -5.05141020e-02 1.29152322e+00 -8.92058685e-02
-3.85190278e-01 7.01605499e-01 -3.33802700e-01 -1.44429386e-01
-2.24146396e-01 1.50151789e-01 -7.45414436e-01 2.00543955e-01
3.37665766e-01 -2.52811220e-02 -3.84956717e-01 4.80151474e-01
-8.68654609e-01 6.56624958e-02 2.99442947e-01 8.48653793e-01
1.25517631e+00 -4.76555377e-01 2.57344872e-01 -1.24435663e+00
6.22743554e-02 -2.07531489e-02 -8.17075968e-01 4.05176967e-01
-9.39147994e-02 7.80924484e-02 3.14527810e-01 -7.30210915e-02
-1.10102212e+00 5.78469813e-01 -1.37501538e-01 -3.57753754e-01
-6.57183409e-01 1.40986284e-02 -4.98349667e-01 -1.33565605e-01
5.46717346e-01 9.29344445e-02 -3.30023557e-01 -2.00092167e-01
1.06261456e+00 5.23634553e-01 9.26376224e-01 -9.04140055e-01
4.27486002e-01 7.62683868e-01 -1.48705229e-01 -8.22232008e-01
-1.28985894e+00 -8.25275838e-01 -1.08417976e+00 -9.60714072e-02
1.71640837e+00 -8.49778712e-01 -3.24621975e-01 1.49226412e-01
-1.34791505e+00 -7.99344122e-01 -5.98954856e-01 -2.69281000e-01
-1.08186805e+00 1.95647687e-01 -3.74454319e-01 -6.60509765e-01
-2.17922926e-01 -1.25772071e+00 1.75483048e+00 2.32148275e-01
-1.56876892e-01 -1.10533059e+00 -2.69743145e-01 7.63332784e-01
-1.07024178e-01 6.22483611e-01 7.27726281e-01 -3.83744031e-01
-8.14561427e-01 -9.24920142e-02 -6.22306228e-01 4.28843319e-01
-2.12545797e-01 -1.49237648e-01 -1.20736778e+00 1.89994916e-01
-4.62399542e-01 -6.32950127e-01 1.12292790e+00 4.68025655e-01
1.39277673e+00 -7.94845670e-02 -3.34954053e-01 7.49511838e-01
1.57524312e+00 -8.15319568e-02 4.59812671e-01 1.21944264e-01
1.05112505e+00 1.04846835e+00 8.20610404e-01 -1.19755283e-01
1.56888306e-01 7.58688867e-01 4.46687967e-01 -8.21662486e-01
-1.94768354e-01 -3.32150757e-01 -1.11562952e-01 -2.27778599e-01
2.37301141e-01 -3.54172677e-01 -1.02051151e+00 9.57685351e-01
-1.88849640e+00 -6.08386993e-01 -5.06676435e-01 1.64673924e+00
8.54582250e-01 8.66366178e-02 2.38034695e-01 -8.32932740e-02
7.96990454e-01 -1.77889727e-02 -5.76287031e-01 -6.65071070e-01
-2.55248994e-01 3.67494226e-01 6.14307165e-01 6.25112176e-01
-1.34338617e+00 1.69698930e+00 5.02925110e+00 8.56979489e-01
-8.97600770e-01 5.72194874e-01 1.09269869e+00 3.61470759e-01
-2.42343158e-01 1.32234648e-01 -8.36920083e-01 4.34548184e-02
4.50925976e-01 2.92979389e-01 8.67015310e-03 1.01215577e+00
1.39798060e-01 -4.43942726e-01 -1.12422252e+00 1.07652938e+00
2.68892258e-01 -1.13964701e+00 -1.56015474e-02 -3.31333905e-01
1.13819134e+00 -4.83335200e-04 -1.13776780e-03 -6.96377233e-02
1.87037840e-01 -8.92211497e-01 1.08351493e+00 2.36037001e-01
7.58692861e-01 -5.81725001e-01 6.80339515e-01 2.75025181e-02
-1.06091547e+00 2.54825592e-01 -1.46857485e-01 4.62018758e-01
4.51531708e-01 3.99719745e-01 -6.65993929e-01 1.58822358e-01
5.62820852e-01 8.38897645e-01 -6.24345183e-01 5.71680427e-01
-6.18430376e-01 7.38123238e-01 -2.20136777e-01 1.38239235e-01
7.47370481e-01 -1.89662755e-01 1.96051806e-01 1.44275725e+00
-2.53537685e-01 -6.36666343e-02 3.46803069e-01 1.45308864e+00
-2.20439360e-01 1.93831816e-01 -3.70596439e-01 -7.40196705e-02
-1.48145393e-01 1.46346843e+00 -1.42904043e+00 -3.17557484e-01
-1.72671646e-01 1.51253831e+00 2.40584224e-01 6.07971132e-01
-8.06268394e-01 -2.69781739e-01 8.89253542e-02 2.35831469e-01
4.14796680e-01 2.83992570e-02 -5.31699121e-01 -8.85396838e-01
-2.41175573e-02 -4.81289715e-01 3.84558856e-01 -1.01179600e+00
-9.47620988e-01 5.42266846e-01 1.07551686e-01 -8.97614539e-01
2.93163788e-02 -6.95843399e-01 -4.48690593e-01 7.04758286e-01
-1.46525490e+00 -1.49937689e+00 -5.03966928e-01 6.06553733e-01
8.05852056e-01 3.68403435e-01 6.17103636e-01 -9.65092257e-02
-6.17930293e-01 3.14656526e-01 -3.70272577e-01 1.84203550e-01
3.55240226e-01 -1.73077750e+00 3.29202294e-01 9.31734681e-01
4.35143679e-01 4.21062022e-01 7.54994750e-01 -3.08229148e-01
-7.73614287e-01 -1.23412752e+00 7.00936377e-01 -5.59254348e-01
5.21342039e-01 -8.04466605e-01 -9.12051916e-01 6.09818161e-01
3.72182906e-01 4.56424206e-01 3.80648077e-01 -1.60946444e-01
-2.23978341e-01 1.41411483e-01 -1.17813361e+00 3.07582736e-01
1.00117683e+00 -7.35689461e-01 -4.57666278e-01 7.38234937e-01
9.01111484e-01 -5.05661428e-01 -5.40236712e-01 8.14926252e-02
-1.26416311e-01 -7.82075226e-01 8.69118929e-01 -4.04464364e-01
4.13493603e-01 -3.98937047e-01 -1.61846414e-01 -7.85016060e-01
6.09870017e-01 -6.87278330e-01 6.63072288e-01 1.43105638e+00
4.99633580e-01 -2.78142929e-01 1.06841969e+00 4.97012019e-01
-1.22532569e-01 -6.40705168e-01 -7.39061058e-01 -5.64198732e-01
7.47938603e-02 -6.77734315e-01 1.48884013e-01 7.53604829e-01
-1.68343037e-01 4.09303963e-01 1.43146291e-01 1.28264539e-02
6.43263876e-01 2.93065280e-01 8.17022741e-01 -7.87922502e-01
-2.00348511e-01 -4.13705170e-01 -5.41784585e-01 -1.24137008e+00
5.52619874e-01 -1.04334593e+00 5.44175088e-01 -1.70721877e+00
1.95808738e-01 -3.40828121e-01 2.75563359e-01 7.50879884e-01
-1.64634719e-01 7.17707694e-01 1.98050514e-01 -1.03968620e-01
-1.10607612e+00 3.40547502e-01 1.12671113e+00 -4.37203765e-01
-1.00427821e-01 -1.91026226e-01 -4.28935260e-01 9.15451944e-01
6.47362649e-01 -4.91410106e-01 -4.37403828e-01 -2.60489106e-01
-5.49498722e-02 -1.03879847e-01 7.03749955e-01 -5.42424500e-01
-1.29119129e-04 1.01933427e-01 -1.76016644e-01 -7.91265786e-01
1.17695279e-01 -7.00645506e-01 -4.72403675e-01 -5.52251190e-03
-5.21148324e-01 -4.30136651e-01 1.69504479e-01 6.24038696e-01
-3.33492965e-01 -5.24174929e-01 9.68946457e-01 -3.02946627e-01
-9.09886599e-01 2.41878591e-02 -2.63119638e-01 5.20430684e-01
1.19329727e+00 -3.78778636e-01 1.34165902e-02 -2.66241491e-01
-1.19492829e+00 3.57000351e-01 5.42114377e-01 1.46148622e-01
5.14370620e-01 -8.13461900e-01 -3.77457112e-01 -1.44507870e-01
5.08723736e-01 4.93968755e-01 1.05687156e-01 9.53209937e-01
-7.18835473e-01 4.31270897e-01 1.49433210e-01 -1.29435432e+00
-1.43586934e+00 5.75444281e-01 7.01901019e-01 -2.01232899e-02
-4.79392558e-01 1.31183505e+00 6.67723060e-01 -5.13431609e-01
1.89118147e-01 -6.86712563e-01 1.00526519e-01 6.70347959e-02
2.08343603e-02 -1.02009848e-01 1.51451137e-02 -1.07359171e+00
-1.90989345e-01 1.08284974e+00 3.37589756e-02 -1.63087070e-01
9.05928552e-01 -2.47451678e-01 -1.67821079e-01 3.60139161e-01
1.33163810e+00 -3.67045194e-01 -1.48379409e+00 -2.86179096e-01
3.65558267e-01 -3.16307813e-01 1.28953144e-01 -5.99686921e-01
-1.17316401e+00 1.19922912e+00 4.14787650e-01 -4.63324338e-02
1.10671163e+00 6.41171575e-01 6.42310798e-01 2.34014660e-01
4.69824642e-01 -1.03248203e+00 3.62865120e-01 2.63464659e-01
6.55220985e-01 -1.33156228e+00 -3.19512814e-01 -8.12810183e-01
-8.33045661e-01 9.07585859e-01 5.39059579e-01 -8.89741853e-02
8.40879008e-02 2.00585544e-01 2.20608607e-01 -4.16554153e-01
-1.58188745e-01 -9.39059496e-01 3.86746198e-01 6.99912488e-01
4.58680511e-01 -1.15928538e-01 -2.22103491e-01 3.60394537e-01
1.26568154e-01 -2.37843439e-01 2.87417412e-01 7.42904305e-01
-4.16718632e-01 -1.09775496e+00 -4.19922650e-01 -4.56959233e-02
-5.65669417e-01 -1.45295352e-01 -7.70179629e-01 7.10999846e-01
3.66975427e-01 8.54222059e-01 1.77486822e-01 1.92674875e-01
4.19906318e-01 2.04540178e-01 4.75018293e-01 -8.83762002e-01
-5.70670843e-01 4.62554783e-01 -1.50735741e-02 -6.88831210e-01
-8.90065253e-01 -6.85546219e-01 -1.78692746e+00 5.36037564e-01
-3.97689283e-01 1.52661428e-01 6.30913138e-01 1.06458616e+00
-5.75070307e-02 5.50466001e-01 1.65406600e-01 -9.47029412e-01
2.38301426e-01 -5.49631834e-01 -4.41104144e-01 6.76836133e-01
1.34536758e-01 -4.13574994e-01 -1.15289137e-01 5.94873488e-01]
|
[9.964070320129395, 1.0950195789337158]
|
934b055e-d3bd-4f43-ad72-65b64405307b
|
contrastive-learning-for-unsupervised-video
| null | null |
http://openaccess.thecvf.com//content/CVPR2022/html/Badamdorj_Contrastive_Learning_for_Unsupervised_Video_Highlight_Detection_CVPR_2022_paper.html
|
http://openaccess.thecvf.com//content/CVPR2022/papers/Badamdorj_Contrastive_Learning_for_Unsupervised_Video_Highlight_Detection_CVPR_2022_paper.pdf
|
Contrastive Learning for Unsupervised Video Highlight Detection
|
Video highlight detection can greatly simplify video browsing, potentially paving the way for a wide range of applications. Existing efforts are mostly fully-supervised, requiring humans to manually identify and label the interesting moments (called highlights) in a video. Recent weakly supervised methods forgo the use of highlight annotations, but typically require extensive efforts in collecting external data such as web-crawled videos for model learning. This observation has inspired us to consider unsupervised highlight detection where neither frame-level nor video-level annotations are available in training. We propose a simple contrastive learning framework for unsupervised highlight detection. Our framework encodes a video into a vector representation by learning to pick video clips that help to distinguish it from other videos via a contrastive objective using dropout noise. This inherently allows our framework to identify video clips corresponding to highlight of the video. Extensive empirical evaluations on three highlight detection benchmarks demonstrate the superior performance of our approach.
|
['Li Cheng', 'Yang Wang', 'Mrigank Rochan', 'Taivanbat Badamdorj']
|
2022-01-01
| null | null | null |
cvpr-2022-1
|
['highlight-detection']
|
['computer-vision']
|
[ 4.57087398e-01 -3.52525771e-01 -5.39247334e-01 -2.82506526e-01
-8.81090164e-01 -7.84575164e-01 5.79717398e-01 4.69890743e-01
-4.54698980e-01 3.95795196e-01 2.42133200e-01 4.45342362e-02
2.07983568e-01 -3.91339570e-01 -9.08834517e-01 -7.69735217e-01
-6.02721393e-01 -3.20396572e-01 5.13927281e-01 2.30317950e-01
4.00395334e-01 2.92156368e-01 -1.61439323e+00 5.65460324e-01
4.50596571e-01 9.83676493e-01 2.83879846e-01 6.77552700e-01
-6.40662014e-02 1.20138979e+00 -4.91326928e-01 -1.73963740e-01
4.50506836e-01 -3.81154209e-01 -5.04905105e-01 5.25477827e-01
7.69770563e-01 -6.13652110e-01 -5.32530487e-01 9.98254240e-01
1.55146390e-01 4.53763694e-01 3.58812690e-01 -1.43791878e+00
-1.60768047e-01 6.86575413e-01 -6.95307851e-01 6.66965187e-01
4.00747150e-01 1.29946202e-01 1.29734409e+00 -9.33433473e-01
9.48572397e-01 7.93065667e-01 6.04516268e-01 1.34751827e-01
-1.16654384e+00 -3.95784229e-01 4.17373657e-01 4.63546216e-01
-1.15546262e+00 -3.40076715e-01 1.23518264e+00 -5.92545152e-01
3.93241704e-01 3.77837896e-01 6.22634292e-01 1.09062386e+00
-3.46562803e-01 1.22096848e+00 1.00188828e+00 -3.93794745e-01
2.58015424e-01 2.91805584e-02 1.18015558e-02 8.81866693e-01
9.08409879e-02 -1.59050733e-01 -1.08037567e+00 -2.14886919e-01
6.75683081e-01 2.96473444e-01 -3.94106001e-01 -6.42948151e-01
-1.26387465e+00 6.83049619e-01 2.38398612e-01 -2.21602693e-02
-4.51232731e-01 1.87843695e-01 8.10758829e-01 4.56056952e-01
5.29273331e-01 3.26539040e-01 -2.76303411e-01 -3.23541492e-01
-1.40988779e+00 2.23463759e-01 5.15155613e-01 9.01945353e-01
8.59991372e-01 1.26339281e-02 -3.57755601e-01 6.28183007e-01
-1.90226942e-01 1.51673242e-01 1.16318293e-01 -1.14681339e+00
3.65030080e-01 5.86956441e-01 1.07183069e-01 -1.09729254e+00
-2.10970733e-02 -1.79196134e-01 -2.92120546e-01 2.58487791e-01
4.74593788e-01 5.20751514e-02 -7.16854811e-01 1.47724819e+00
2.47690395e-01 5.66105485e-01 -4.32368457e-01 1.03107595e+00
5.69823384e-01 6.03698373e-01 5.71701899e-02 -2.52004445e-01
1.12421119e+00 -1.10337555e+00 -4.94857699e-01 1.59981146e-01
5.82204163e-01 -8.60181451e-01 1.33055627e+00 6.05836928e-01
-1.05086589e+00 -4.24281210e-01 -8.92998457e-01 -4.09124233e-02
-2.49580964e-01 1.15569830e-01 6.50058746e-01 4.42574382e-01
-7.16071963e-01 5.87057412e-01 -7.68429339e-01 -2.08265707e-01
5.83751559e-01 8.08224678e-02 -4.85594243e-01 -7.82674700e-02
-6.49737597e-01 2.19639927e-01 3.31517965e-01 -1.96680620e-01
-1.20461166e+00 -7.15285182e-01 -8.13603163e-01 9.86196771e-02
9.96663392e-01 -1.76576689e-01 1.16761410e+00 -1.42603779e+00
-1.10508811e+00 9.55927968e-01 -2.65498936e-01 -5.47000170e-01
7.12496400e-01 -7.79880226e-01 -1.40897825e-01 7.42047250e-01
6.35454897e-03 6.76846027e-01 1.24371970e+00 -1.30339956e+00
-9.78094876e-01 2.61317521e-01 4.73378658e-01 -4.77226861e-02
-7.56640077e-01 4.01088744e-01 -9.21624601e-01 -1.08060956e+00
-2.19103739e-01 -8.15376878e-01 2.37263693e-03 3.39928985e-01
-5.20374477e-01 -1.12198457e-01 1.23110366e+00 -5.67608535e-01
1.40470195e+00 -2.10849690e+00 -1.06201112e-01 2.02630505e-01
4.74171042e-01 5.99351488e-02 -1.70574471e-01 5.25476754e-01
1.18743733e-01 -7.56312683e-02 -2.00563986e-02 -2.37713918e-01
-1.56075194e-01 -1.38892293e-01 -4.30620283e-01 5.50581932e-01
4.55445617e-01 4.61078823e-01 -1.14773798e+00 -7.75925040e-01
2.46005848e-01 4.76548731e-01 -7.44395256e-01 2.74338990e-01
-3.12410504e-01 3.14515710e-01 -3.00501853e-01 8.54666233e-01
3.94782841e-01 -3.07006449e-01 7.89684206e-02 -2.67497808e-01
-2.04927802e-01 9.86978635e-02 -1.17380226e+00 1.56459916e+00
-2.02900283e-02 1.13572586e+00 -1.42771482e-01 -9.13876235e-01
5.33377171e-01 5.85963540e-02 7.57720709e-01 -3.76975149e-01
-3.47282626e-02 -8.70574489e-02 -2.28850231e-01 -7.83810139e-01
6.48834705e-01 2.75441974e-01 1.61522999e-01 3.01626205e-01
2.34430283e-02 4.02326286e-01 5.77568889e-01 4.22520876e-01
1.34516346e+00 3.92495841e-01 -2.53869612e-02 -7.66730905e-02
4.03379261e-01 -3.31242681e-02 6.59366071e-01 8.95935535e-01
-3.04654390e-01 5.77163517e-01 7.39433110e-01 -3.89079154e-01
-1.03102088e+00 -9.53568935e-01 2.75360852e-01 1.66011286e+00
1.18888438e-01 -8.07309091e-01 -7.78663456e-01 -9.74292874e-01
-1.01655811e-01 1.25988901e-01 -6.52448416e-01 1.68728940e-02
-8.22297871e-01 -2.04003789e-02 2.77226299e-01 4.48378891e-01
1.68450236e-01 -8.77820611e-01 -9.91240144e-01 -6.11958746e-03
-1.08697847e-01 -1.28222108e+00 -7.52600491e-01 2.59036362e-01
-8.67842376e-01 -1.29051363e+00 -8.27349782e-01 -9.10715759e-01
7.56054640e-01 6.50414765e-01 1.04558742e+00 3.20004284e-01
-4.29662913e-01 5.84465623e-01 -5.13427794e-01 -3.36692184e-02
-5.34817576e-02 -1.02902696e-01 9.34234634e-03 1.78387508e-01
3.05649400e-01 -3.20014805e-01 -6.92123711e-01 2.55007774e-01
-1.06482387e+00 1.65762492e-02 3.57725292e-01 8.08262169e-01
7.49713004e-01 9.22803730e-02 1.52600735e-01 -9.93061662e-01
3.59390467e-01 -4.07986253e-01 -4.73628908e-01 1.06151007e-01
-9.88609642e-02 -1.99056953e-01 6.84329927e-01 -6.01776361e-01
-7.36670971e-01 2.33191401e-01 5.26600838e-01 -9.09071565e-01
-1.27506942e-01 4.70370054e-01 5.95059292e-03 -1.78195074e-01
4.24421430e-01 9.25782174e-02 -3.59140813e-01 -4.84886885e-01
3.97336811e-01 2.05189213e-01 7.46976852e-01 -5.12366772e-01
9.01520610e-01 7.02239990e-01 -1.07814737e-01 -9.67172742e-01
-8.71052980e-01 -7.94036686e-01 -6.50164425e-01 -6.47237122e-01
4.44504797e-01 -9.81750250e-01 -5.78377485e-01 1.06991336e-01
-6.22334898e-01 -4.39627469e-01 -2.00894609e-01 4.47993875e-01
-5.76836824e-01 6.88218892e-01 -5.78028679e-01 -8.05916429e-01
-2.33145822e-02 -9.05886233e-01 1.09199679e+00 2.01040179e-01
-2.93138593e-01 -8.87408912e-01 -1.26318380e-01 8.66360888e-02
1.57249406e-01 5.12570500e-01 6.05504453e-01 -5.02070844e-01
-7.81969726e-01 -4.26729381e-01 -2.28431404e-01 2.53203958e-01
1.34209409e-01 3.35705906e-01 -1.05523229e+00 -3.06723624e-01
-3.64755362e-01 -3.71176541e-01 1.12953007e+00 3.43098789e-01
1.42391682e+00 -3.42632055e-01 -3.04563671e-01 6.07193053e-01
1.38968587e+00 -1.45101875e-01 1.99312866e-01 5.51226616e-01
7.82615125e-01 5.66931605e-01 8.42212439e-01 6.86612904e-01
-1.39589310e-02 6.63172901e-01 3.49820822e-01 -1.93489581e-01
-9.09390450e-02 -3.97593558e-01 5.97940505e-01 4.16056603e-01
-2.73219079e-01 -9.49076042e-02 -6.50051832e-01 5.12351096e-01
-1.81845355e+00 -1.34294140e+00 1.51412815e-01 2.23953867e+00
7.59554923e-01 4.76020604e-01 5.35755992e-01 3.17060888e-01
7.96860516e-01 4.29770201e-01 -4.80613410e-01 1.08748414e-01
-1.34131610e-01 -4.13758308e-02 4.34800774e-01 1.06992880e-02
-1.59510839e+00 9.21249330e-01 6.01852083e+00 5.49848616e-01
-1.24761379e+00 -4.94115129e-02 5.97514331e-01 -6.00900292e-01
-6.15312122e-02 1.95991751e-02 -3.16997647e-01 6.97083414e-01
5.80821872e-01 -1.67726986e-02 2.38932535e-01 1.07342684e+00
6.09117091e-01 -2.89618909e-01 -1.22303617e+00 1.16087508e+00
2.13780388e-01 -1.50024676e+00 -1.54810548e-02 -2.41549715e-01
9.04868722e-01 -1.06359750e-01 1.56326771e-01 1.37129232e-01
-1.85888946e-01 -5.78613997e-01 8.43891263e-01 5.12845218e-01
5.67011893e-01 -6.80902839e-01 2.45430633e-01 -1.19436078e-01
-1.34671390e+00 -2.10188612e-01 -2.14825153e-01 -1.05743818e-01
1.00262262e-01 5.35819292e-01 -6.13957584e-01 5.74635789e-02
8.46030295e-01 9.37071562e-01 -6.02275848e-01 1.55147433e+00
-1.19543329e-01 1.02908015e+00 -2.22669929e-01 1.81502685e-01
3.78302991e-01 4.03570421e-02 6.05099797e-01 1.56746471e+00
2.87285689e-02 -3.17706198e-01 6.42911673e-01 4.61344898e-01
-4.31596518e-01 2.71274656e-01 -4.70537245e-01 -2.59321541e-01
3.71432126e-01 1.34377396e+00 -1.18565142e+00 -5.34349680e-01
-6.33903623e-01 1.10182965e+00 1.48465440e-01 3.88981640e-01
-8.89592648e-01 -4.18411583e-01 3.97604525e-01 2.97865808e-01
5.64595222e-01 -3.34104508e-01 1.36172429e-01 -1.21812844e+00
2.92078048e-01 -9.39319074e-01 4.60828930e-01 -5.70746183e-01
-9.33533967e-01 2.77368516e-01 -1.35939807e-01 -1.56356764e+00
-5.43259270e-02 -4.96692628e-01 -8.61073971e-01 1.22457325e-01
-1.61281109e+00 -9.15954411e-01 -5.48294902e-01 6.51706815e-01
8.81453693e-01 -2.29330957e-02 2.87148148e-01 4.46509361e-01
-5.38035631e-01 5.24359345e-01 2.11700007e-01 3.59754086e-01
1.06444156e+00 -1.30803180e+00 -1.60407493e-04 1.03366232e+00
5.51844418e-01 6.16141379e-01 9.08609867e-01 -6.10839069e-01
-1.53399372e+00 -9.33540881e-01 4.45926189e-01 -3.86057198e-01
7.87563980e-01 -3.48225921e-01 -9.06641245e-01 4.65984523e-01
2.09613234e-01 2.31550336e-01 7.81078041e-01 1.19775727e-01
-6.04278207e-01 -4.89012189e-02 -6.52140617e-01 6.04043067e-01
9.70854044e-01 -8.61655831e-01 -3.39221656e-01 3.63795131e-01
3.65875840e-01 -1.32387593e-01 -5.07444024e-01 1.55008987e-01
5.75754523e-01 -9.93996143e-01 8.32674146e-01 -6.81164622e-01
6.62902176e-01 -3.84300619e-01 2.30798181e-02 -7.81397164e-01
6.52097762e-02 -9.77979660e-01 -5.58100045e-01 1.26839662e+00
-4.85709943e-02 2.33639807e-01 9.34368372e-01 4.12703633e-01
-4.32805642e-02 -6.77207351e-01 -4.80785191e-01 -8.44451427e-01
-6.06289089e-01 -3.84176373e-01 1.74976364e-02 1.12389123e+00
1.02776848e-01 -1.16966791e-01 -7.01895237e-01 1.60205469e-01
6.63195372e-01 1.03900567e-01 9.21996415e-01 -1.22142136e+00
-2.55398661e-01 -6.25769436e-01 -5.53203046e-01 -1.06679475e+00
5.82299568e-02 -6.80286288e-01 1.58494696e-01 -1.08545852e+00
5.85671425e-01 -1.02761544e-01 -6.85639679e-01 4.33564901e-01
-3.05271775e-01 4.74138528e-01 2.74682522e-01 4.08349782e-01
-1.22182620e+00 7.65278116e-02 6.95663631e-01 -1.74797669e-01
-1.49516433e-01 -2.10852996e-01 -3.59466344e-01 9.51681495e-01
6.33092821e-01 -4.98750001e-01 -3.15467119e-01 -1.14805788e-01
1.53296694e-01 -2.65175849e-01 5.49874485e-01 -1.13898790e+00
6.49906397e-02 -2.14489251e-01 4.59977537e-01 -4.86092150e-01
2.00691625e-01 -7.81655073e-01 -3.76469672e-01 1.66254058e-01
-6.83407366e-01 4.05811816e-02 1.29413486e-01 8.05789709e-01
-3.80889267e-01 -2.39294246e-01 6.05194092e-01 8.13184753e-02
-1.14638245e+00 2.96745539e-01 -3.78174990e-01 4.72153723e-02
1.03953362e+00 -2.85212338e-01 -1.85285270e-01 -4.31713820e-01
-5.32125592e-01 4.22579609e-02 7.35452294e-01 3.99308711e-01
5.55983722e-01 -9.99701858e-01 -4.11177993e-01 4.93454449e-02
3.08652312e-01 -5.09814203e-01 2.31933549e-01 8.92224610e-01
-5.41769147e-01 -2.27389839e-02 -1.94381982e-01 -5.21358311e-01
-1.53916359e+00 8.16177607e-01 -1.95821375e-01 3.29359956e-02
-9.50604618e-01 8.08151007e-01 1.76951513e-01 5.43316543e-01
6.38639450e-01 -3.29901874e-01 2.38903798e-03 3.24952871e-01
8.86420488e-01 4.11319971e-01 -2.96420932e-01 -5.45324445e-01
-1.74835518e-01 2.87466377e-01 -2.76188076e-01 1.93691105e-01
1.24586225e+00 -7.98135176e-02 3.25703382e-01 4.46802109e-01
1.40526688e+00 3.72586787e-01 -1.72306514e+00 -3.88352990e-01
4.48462158e-01 -6.93466961e-01 5.07201403e-02 -3.49918455e-01
-1.04388082e+00 7.49182224e-01 5.28311908e-01 3.01520586e-01
1.19833076e+00 -5.55791222e-02 7.93732703e-01 7.00350404e-01
2.15240702e-01 -1.39588964e+00 5.58604002e-01 9.34941247e-02
4.95054305e-01 -1.51465106e+00 1.20440364e-01 -3.93279105e-01
-6.54397905e-01 1.23111641e+00 3.25803131e-01 -4.18235660e-01
3.54087800e-01 2.88575798e-01 1.17442019e-01 -1.15966327e-01
-7.14040637e-01 -3.67595643e-01 3.89065295e-01 4.17001277e-01
4.65743572e-01 -2.24823058e-01 -1.32406473e-01 2.70278692e-01
3.00500065e-01 -8.48842338e-02 6.10993147e-01 1.23192644e+00
-7.48490632e-01 -9.00081456e-01 -1.86504304e-01 5.83146214e-01
-7.62873888e-01 -9.34894532e-02 -4.38595057e-01 6.53733194e-01
-5.75141348e-02 7.20812380e-01 8.09546784e-02 -2.88575798e-01
1.98515914e-02 1.64653137e-01 2.50733137e-01 -4.65172470e-01
-5.04896343e-01 4.64441270e-01 -5.69086298e-02 -8.10868800e-01
-6.79097116e-01 -7.49666214e-01 -1.05006182e+00 -5.67701994e-04
-2.61946410e-01 1.83753759e-01 2.98938185e-01 5.97066402e-01
2.21853226e-01 4.02984381e-01 7.43159056e-01 -1.00765812e+00
-8.30134898e-02 -4.93689954e-01 -5.81326365e-01 8.18262815e-01
5.29647350e-01 -8.13810408e-01 -3.48235548e-01 7.37480462e-01]
|
[10.093179702758789, 0.434783935546875]
|
52250e2a-05b6-45d3-83c5-2a55fba69d01
|
a-unified-benchmark-for-the-unknown-detection
|
2112.00337
| null |
https://arxiv.org/abs/2112.00337v1
|
https://arxiv.org/pdf/2112.00337v1.pdf
|
A Unified Benchmark for the Unknown Detection Capability of Deep Neural Networks
|
Deep neural networks have achieved outstanding performance over various tasks, but they have a critical issue: over-confident predictions even for completely unknown samples. Many studies have been proposed to successfully filter out these unknown samples, but they only considered narrow and specific tasks, referred to as misclassification detection, open-set recognition, or out-of-distribution detection. In this work, we argue that these tasks should be treated as fundamentally an identical problem because an ideal model should possess detection capability for all those tasks. Therefore, we introduce the unknown detection task, an integration of previous individual tasks, for a rigorous examination of the detection capability of deep neural networks on a wide spectrum of unknown samples. To this end, unified benchmark datasets on different scales were constructed and the unknown detection capabilities of existing popular methods were subject to comparison. We found that Deep Ensemble consistently outperforms the other approaches in detecting unknowns; however, all methods are only successful for a specific type of unknown. The reproducible code and benchmark datasets are available at https://github.com/daintlab/unknown-detection-benchmarks .
|
['Sangheum Hwang', 'Jiin Koo', 'Jihyo Kim']
|
2021-12-01
| null | null | null | null |
['open-set-learning']
|
['miscellaneous']
|
[ 6.22003153e-03 -2.50149578e-01 9.78094265e-02 -6.44585907e-01
-9.08905029e-01 -6.92426980e-01 6.22087777e-01 -1.87504128e-01
-1.88789740e-01 1.08832073e+00 -4.70258534e-01 -2.40478769e-01
-1.80194512e-01 -6.16306841e-01 -6.38059616e-01 -6.59640074e-01
1.10031091e-01 5.90196609e-01 2.09775135e-01 1.98811311e-02
2.66858965e-01 4.26474690e-01 -1.93206954e+00 2.96611100e-01
8.58851492e-01 1.19695055e+00 -1.35340750e-01 4.45335776e-01
-7.09377378e-02 4.87773955e-01 -9.08354223e-01 -6.54117525e-01
3.12965423e-01 -1.87601060e-01 -3.78565669e-01 -3.73850614e-01
8.90320122e-01 -5.15110791e-01 -2.31556028e-01 1.11964476e+00
8.23954046e-01 -1.98966801e-01 9.33022380e-01 -1.56891620e+00
-7.71507502e-01 4.57563937e-01 -4.65992212e-01 3.96435976e-01
-4.21088841e-03 -3.67341191e-02 1.00630689e+00 -1.28140461e+00
2.27093384e-01 1.01123846e+00 1.13812160e+00 5.10105371e-01
-1.00337756e+00 -1.13110864e+00 -1.08219171e-02 2.63265729e-01
-1.56151211e+00 -5.50858796e-01 3.54671866e-01 -5.92067122e-01
6.83025777e-01 3.76244277e-01 -4.62372825e-02 1.52473807e+00
3.10065478e-01 6.68723524e-01 1.01152289e+00 -6.88899234e-02
1.32306606e-01 4.14355963e-01 6.26255929e-01 3.85371238e-01
7.80514002e-01 3.19816083e-01 -2.93406487e-01 -2.75684267e-01
2.79156804e-01 1.91660300e-01 -3.98393393e-01 -8.17932859e-02
-8.35728705e-01 7.90147662e-01 6.05028421e-02 9.31100547e-02
-1.23778485e-01 -3.42180490e-01 4.67755497e-01 4.81036365e-01
7.41305828e-01 5.93547165e-01 -9.16978240e-01 1.65440291e-01
-7.77874708e-01 3.40553731e-01 1.05045760e+00 7.70944178e-01
5.46449482e-01 1.94579706e-01 -2.73559749e-01 1.06379294e+00
-1.22783639e-01 3.94016325e-01 5.04570663e-01 -6.68007255e-01
3.12142670e-01 5.37221491e-01 3.49956185e-01 -1.02479744e+00
-4.87361372e-01 -9.26353335e-01 -1.14674270e+00 1.05040260e-01
7.62998998e-01 -4.81308609e-01 -1.05660558e+00 1.57721090e+00
2.10452422e-01 3.26830685e-01 4.58730049e-02 7.77167559e-01
1.11396897e+00 3.72221529e-01 -1.59853160e-01 -1.22488355e-02
1.12951112e+00 -6.03035092e-01 -6.84377789e-01 -2.87657082e-01
2.59175628e-01 -6.77905440e-01 7.52070785e-01 5.97919285e-01
-5.17560720e-01 -6.54199243e-01 -1.18050694e+00 2.89365768e-01
-7.70791471e-01 4.80228037e-01 5.23327887e-01 7.11124361e-01
-6.97383225e-01 7.05828249e-01 -5.53119421e-01 -1.89258724e-01
6.35321558e-01 2.71738380e-01 -2.22050473e-01 -4.88977320e-02
-1.22854948e+00 8.49154234e-01 3.08815032e-01 3.24433327e-01
-1.20361233e+00 -6.72104895e-01 -5.85813761e-01 1.28487304e-01
5.05810916e-01 -4.78707463e-01 1.31373262e+00 -9.18542504e-01
-8.98757160e-01 8.34990203e-01 7.01966509e-02 -5.56934476e-01
7.68244147e-01 -4.97652292e-01 -6.41092420e-01 -5.76989055e-01
1.46435916e-01 1.59887463e-01 8.65067780e-01 -1.28730631e+00
-7.96370745e-01 -5.07187188e-01 -2.01332539e-01 -2.92415857e-01
-2.71561891e-01 8.10480267e-02 -2.76735798e-03 -6.44892633e-01
2.28644341e-01 -7.37270832e-01 1.02364115e-01 -1.22961663e-01
-7.00644076e-01 -4.23683137e-01 9.35366035e-01 -4.50265378e-01
1.14105189e+00 -2.00787616e+00 -4.55639660e-01 -2.69914299e-01
6.16682053e-01 4.66949612e-01 -1.23301871e-01 2.90225208e-01
-3.24890316e-01 1.67274952e-01 -3.05528283e-01 -4.74174470e-01
1.79877818e-01 1.07237011e-01 -7.29722917e-01 7.39701509e-01
3.72347325e-01 5.94354272e-01 -6.57795548e-01 -4.99941520e-02
-3.55948787e-03 5.40851295e-01 9.57345404e-03 2.03351438e-01
-1.02799900e-01 6.46543428e-02 -2.53786296e-01 1.10174298e+00
9.44726944e-01 -3.16563874e-01 -9.36514586e-02 -1.60708159e-01
3.60366367e-02 -6.97612911e-02 -1.34665477e+00 6.60455346e-01
7.04727918e-02 7.94700682e-01 1.09779998e-03 -1.11400175e+00
1.11316526e+00 6.92320615e-02 2.15265810e-01 -4.16270882e-01
1.97202116e-01 5.34644425e-01 1.68026984e-01 -3.04831386e-01
6.04637742e-01 9.74753946e-02 6.81067556e-02 2.40046397e-01
1.83992982e-01 4.77149218e-01 9.53174606e-02 -3.61107178e-02
1.11215281e+00 -1.60868112e-02 2.80480683e-01 -1.13316409e-01
2.78542787e-01 -5.04562740e-06 9.46596980e-01 1.37610042e+00
-5.51499546e-01 9.26893532e-01 3.65241170e-01 -6.96298838e-01
-8.60273123e-01 -1.13615406e+00 -6.57008767e-01 1.16374886e+00
2.33790874e-01 -1.00129172e-01 -4.54914272e-01 -8.30946982e-01
4.21740204e-01 4.41018075e-01 -7.20956922e-01 2.11647954e-02
-2.47590706e-01 -1.33607900e+00 8.09104264e-01 5.28609753e-01
3.34999055e-01 -8.41738164e-01 -1.35146976e-01 1.27533972e-01
5.22082150e-02 -1.05346608e+00 3.24564353e-02 3.06741834e-01
-5.16367733e-01 -1.53490770e+00 -7.67984450e-01 -6.19354188e-01
4.94545251e-01 2.50934750e-01 1.40761566e+00 2.56170351e-02
-5.64580441e-01 -4.28859629e-02 -2.58424729e-01 -8.72128606e-01
-3.85915637e-01 -6.93708938e-03 3.86288613e-01 7.79903978e-02
6.17755532e-01 -4.13271785e-01 -5.79010487e-01 7.65021265e-01
-7.42090285e-01 -5.26479602e-01 6.63528562e-01 1.23172164e+00
4.20191348e-01 1.06073909e-01 9.98159289e-01 -1.02482438e+00
7.34675348e-01 -8.38575125e-01 -6.68235183e-01 2.74238288e-01
-5.90753496e-01 -1.54303059e-01 8.29505444e-01 -5.07068574e-01
-9.53951120e-01 -1.33284340e-02 -2.60438055e-01 -4.16819632e-01
-5.62426150e-01 2.54191309e-01 -1.24028459e-01 -1.58035266e-03
8.82327795e-01 3.22923273e-01 -1.76266670e-01 -6.82181656e-01
-1.17205255e-01 9.41500425e-01 5.07697105e-01 -4.12842065e-01
6.97981894e-01 3.69642347e-01 -2.44929969e-01 -7.25146830e-01
-1.27425790e+00 -4.31740314e-01 -5.71314812e-01 -2.21811868e-02
2.34366506e-01 -9.93402183e-01 -3.98003072e-01 9.48472559e-01
-1.12588429e+00 -6.67298809e-02 1.14237376e-01 2.93500602e-01
3.72114126e-03 4.42690253e-01 -4.37571108e-01 -1.09735906e+00
-4.47830826e-01 -9.58833992e-01 1.04982066e+00 2.14383736e-01
-7.33048022e-02 -6.27726555e-01 -5.82926348e-02 2.38930225e-01
5.49098492e-01 4.23668832e-01 3.83010149e-01 -1.40511668e+00
-3.07528883e-01 -5.30446827e-01 -5.00541568e-01 7.17522442e-01
2.51587778e-02 2.05749303e-01 -1.44953406e+00 -4.56817418e-01
-1.20059304e-01 -4.97002065e-01 1.26156318e+00 2.90579736e-01
1.44588780e+00 -1.39358595e-01 -5.84477365e-01 6.70142889e-01
1.39087272e+00 1.00014150e-01 4.99877185e-01 2.23054901e-01
4.29550231e-01 3.76323372e-01 6.03094339e-01 5.67590714e-01
1.55854613e-01 3.22256446e-01 5.52622259e-01 -3.85047421e-02
1.43164858e-01 2.47682706e-01 1.25444800e-01 2.60887772e-01
8.49251524e-02 -6.00918174e-01 -9.83981252e-01 5.12338698e-01
-1.77225518e+00 -1.01330721e+00 -3.59555125e-01 2.21685934e+00
5.47128141e-01 2.73776025e-01 -2.25342344e-03 2.16641381e-01
9.47375536e-01 -9.85145718e-02 -9.80988801e-01 1.61033630e-01
-3.60145122e-01 -9.38444510e-02 4.43268538e-01 -1.14674412e-01
-1.44432712e+00 3.94610018e-01 6.71437883e+00 9.80856597e-01
-9.64136362e-01 1.47150308e-01 9.67007577e-01 1.80940479e-01
3.04331928e-01 -4.06471550e-01 -1.34439385e+00 7.16288090e-01
8.18804026e-01 1.60347015e-01 -1.97406471e-01 1.07918417e+00
-1.91961884e-01 2.81574670e-02 -1.12335587e+00 9.20651674e-01
1.25882968e-01 -1.15673947e+00 -1.89721614e-01 -1.39966056e-01
6.78330362e-01 4.18321282e-01 5.56090511e-02 7.72461295e-01
1.76483080e-01 -9.84696209e-01 3.75071406e-01 5.73305488e-01
6.96392357e-01 -5.72581172e-01 1.21827805e+00 4.19094384e-01
-6.79143846e-01 -2.12846816e-01 -5.94087958e-01 -1.92778885e-01
-2.81148136e-01 9.28854764e-01 -7.80583918e-01 3.75757277e-01
9.76827264e-01 6.46712065e-01 -7.28656471e-01 1.31619334e+00
1.05372876e-01 7.71289408e-01 -3.16827834e-01 -1.00817662e-02
-1.39638230e-01 9.30354297e-02 4.51238096e-01 1.03616214e+00
5.32373309e-01 -2.03769252e-01 -4.94646281e-02 1.05824351e+00
-2.22286999e-01 -2.67508507e-01 -6.12673521e-01 1.39971105e-02
5.76092541e-01 1.24946809e+00 -4.78533715e-01 -2.88507521e-01
-3.94725621e-01 7.20176518e-01 3.94190729e-01 2.30350748e-01
-1.15528691e+00 -5.95082998e-01 1.03322029e+00 -4.25385982e-02
3.14688414e-01 8.55952650e-02 -2.28973493e-01 -1.32862663e+00
1.27741531e-01 -8.56655061e-01 6.83181703e-01 -5.10802388e-01
-2.02393341e+00 5.81305265e-01 -4.43067461e-01 -1.32272828e+00
1.31827787e-01 -1.12759435e+00 -5.23559451e-01 7.00868428e-01
-1.54344404e+00 -7.72942245e-01 -6.54505968e-01 2.34277964e-01
5.36826730e-01 -4.12440181e-01 8.01213205e-01 5.55997670e-01
-1.09696651e+00 9.20711040e-01 7.49111235e-01 3.55697215e-01
1.13990331e+00 -1.27615392e+00 3.59611034e-01 8.10057580e-01
-3.11553907e-02 4.92985576e-01 7.57352471e-01 -5.66263080e-01
-1.22473860e+00 -1.34159112e+00 4.69387352e-01 -8.60311389e-01
5.89904964e-01 -4.77014869e-01 -1.25033116e+00 6.31139636e-01
-1.75602958e-01 2.45105758e-01 7.34845996e-01 2.44010642e-01
-3.51356268e-01 -2.61375338e-01 -1.13609278e+00 1.52281716e-01
9.23564911e-01 -1.32195890e-01 -4.84952956e-01 5.81768036e-01
5.13230860e-01 -6.53644800e-01 -8.19931209e-01 7.67284691e-01
7.82100558e-01 -1.28323376e+00 1.03168523e+00 -7.93472886e-01
4.18697715e-01 -3.03876787e-01 -2.40215942e-01 -1.10578716e+00
-2.30308220e-01 -1.64389387e-01 -3.33349049e-01 1.38701618e+00
6.37379348e-01 -1.11248195e+00 8.08625281e-01 3.82004023e-01
-3.30514163e-01 -7.40803778e-01 -8.36475849e-01 -9.01052892e-01
6.21737987e-02 -4.02971238e-01 6.49304211e-01 1.20591509e+00
-7.37994313e-01 2.45356724e-01 -6.27972782e-01 5.75764358e-01
8.15805852e-01 1.99131593e-01 8.48807335e-01 -1.63401306e+00
-1.98806062e-01 -4.08041149e-01 -3.54234308e-01 -6.96485996e-01
1.79971173e-01 -5.88199139e-01 1.37608498e-01 -1.09716809e+00
2.80473083e-01 -3.86673123e-01 -5.68465590e-01 4.56091344e-01
-3.51554573e-01 5.10902464e-01 -2.60088742e-01 4.35131907e-01
-6.94655836e-01 4.16860938e-01 8.01120102e-01 -2.87609488e-01
1.03705287e-01 4.13189560e-01 -9.00855720e-01 6.91363394e-01
1.02346647e+00 -5.21261156e-01 -2.07686778e-02 -2.75453001e-01
-1.27455387e-02 -2.83364892e-01 5.70416689e-01 -1.24510348e+00
7.43208081e-02 -1.68165490e-02 7.64007509e-01 -8.10965121e-01
3.54224533e-01 -5.31957269e-01 8.36135447e-02 3.79288584e-01
-1.51176244e-01 -1.58701733e-01 3.71114492e-01 7.35558629e-01
-2.30711296e-01 -9.94085148e-02 7.26431847e-01 -2.88238134e-02
-9.84889984e-01 3.61991733e-01 -1.24485880e-01 2.03720897e-01
1.08620167e+00 -2.26748437e-01 -8.05867076e-01 -2.84862101e-01
-7.37411976e-01 1.71344638e-01 8.81592706e-02 5.23660421e-01
5.58801472e-01 -8.93825173e-01 -1.00292444e+00 2.64221281e-01
2.90548742e-01 -1.07569963e-01 2.18449548e-01 6.96945131e-01
-1.81173787e-01 3.43027472e-01 -1.19853936e-01 -7.98773468e-01
-1.19257891e+00 3.72582614e-01 7.74409592e-01 7.31673371e-03
-1.42773569e-01 8.72195721e-01 2.95947194e-01 -8.64974141e-01
5.66056967e-01 -2.38582417e-01 -1.43578678e-01 7.85266906e-02
6.37729406e-01 6.23706341e-01 2.90800452e-01 -3.89104217e-01
-3.91310126e-01 1.14714451e-01 -1.08625762e-01 9.61302757e-01
1.46670187e+00 2.09756438e-02 2.39044148e-02 3.72815400e-01
8.95554483e-01 -3.70214045e-01 -1.09417140e+00 -1.11705191e-01
1.81155233e-03 -3.75349551e-01 -2.03350365e-01 -1.18622792e+00
-9.53951955e-01 7.33198047e-01 8.22085083e-01 5.87408721e-01
9.08692300e-01 -1.93321735e-01 5.16778648e-01 5.12937844e-01
1.99864507e-01 -8.67028892e-01 -2.22717986e-01 6.92417204e-01
8.31045210e-01 -1.82882857e+00 5.20670451e-02 -3.39633465e-01
-2.35811472e-01 1.15077150e+00 1.13650012e+00 3.11684962e-02
6.02681696e-01 3.53387952e-01 7.78120533e-02 -9.10792276e-02
-7.14953125e-01 -6.26621321e-02 3.85149002e-01 5.43367147e-01
5.21866381e-01 3.36942747e-02 -1.50668740e-01 1.03056479e+00
4.43897955e-02 -2.03208812e-02 5.42212963e-01 5.49014807e-01
-3.81203324e-01 -7.16041684e-01 -5.99874258e-01 1.11149633e+00
-7.20744014e-01 6.17705751e-03 -6.84427917e-01 9.01047707e-01
3.23021591e-01 8.16645682e-01 4.83972616e-02 -5.49149334e-01
3.38827282e-01 2.46590432e-02 -4.47979271e-02 -5.11053920e-01
-3.07015091e-01 -5.92327893e-01 1.62532717e-01 -4.30095404e-01
-1.03151724e-01 -4.37956661e-01 -5.66965878e-01 -2.46084154e-01
-6.82123125e-01 2.11054515e-02 4.96979177e-01 8.44849050e-01
6.45611525e-01 4.16671574e-01 3.66491318e-01 -7.53527403e-01
-1.14530182e+00 -1.19594324e+00 -8.23685884e-01 3.59540492e-01
5.50264180e-01 -9.84673023e-01 -8.80474448e-01 -3.81442934e-01]
|
[9.406068801879883, 2.9661829471588135]
|
74888053-663f-494f-acb8-d6a7c7adc725
|
semi-supervised-intent-discovery-with
| null | null |
https://aclanthology.org/2021.nlp4convai-1.12
|
https://aclanthology.org/2021.nlp4convai-1.12.pdf
|
Semi-supervised Intent Discovery with Contrastive Learning
|
User intent discovery is a key step in developing a Natural Language Understanding (NLU) module at the core of any modern Conversational AI system. Typically, human experts review a representative sample of user input data to discover new intents, which is subjective, costly, and error-prone. In this work, we aim to assist the NLU developers by presenting a novel method for discovering new intents at scale given a corpus of utterances. Our method utilizes supervised contrastive learning to leverage information from a domain-relevant, already labeled dataset and identifies new intents in the corpus at hand using unsupervised K-means clustering. Our method outperforms the state-of-the-art by a large margin up to 2% and 13% on two benchmark datasets, measured by clustering accuracy. Furthermore, we apply our method on a large dataset from the travel domain to demonstrate its effectiveness on a real-world use case.
|
['Mani Najmabadi', 'Yao Zhang', 'Yinge Sun', 'Xiang Shen']
| null | null | null | null |
emnlp-nlp4convai-2021-11
|
['intent-discovery']
|
['natural-language-processing']
|
[ 3.41167212e-01 2.39605784e-01 -2.75156885e-01 -6.21052980e-01
-8.03720713e-01 -6.94185734e-01 6.84192777e-01 2.90197462e-01
-4.00445074e-01 4.24644709e-01 4.67133492e-01 -1.33806333e-01
2.47722208e-01 -3.80043000e-01 -3.49509478e-01 -5.18994443e-02
1.28393890e-02 7.39222586e-01 5.86701781e-02 -4.04738218e-01
5.28402984e-01 -1.65646061e-01 -1.56395161e+00 4.34425741e-01
1.05584335e+00 6.51153088e-01 3.68421316e-01 5.77084005e-01
-2.31146142e-01 9.80304241e-01 -3.53997022e-01 -2.70415783e-01
1.22350775e-01 -3.63163710e-01 -1.29551256e+00 2.10709110e-01
3.57727073e-02 -3.82552177e-01 2.00969025e-01 6.50202930e-01
7.78099746e-02 4.38458413e-01 5.96291184e-01 -1.24692523e+00
-3.71047735e-01 8.52380693e-01 -1.13310516e-01 -1.02693941e-02
8.55462611e-01 1.77497074e-01 1.30619633e+00 -8.62443864e-01
6.62563741e-01 8.68577361e-01 4.19760078e-01 8.35149705e-01
-1.04089320e+00 -3.97650629e-01 1.56893745e-01 5.36095053e-02
-1.11068439e+00 -6.55438721e-01 1.04413307e+00 -5.87804019e-01
1.09473336e+00 3.05157572e-01 6.75093681e-02 1.10292780e+00
-2.80641228e-01 1.19325888e+00 8.67961049e-01 -5.16273320e-01
6.48550391e-01 4.35849100e-01 7.70507038e-01 5.56878686e-01
-3.01765740e-01 -4.94779259e-01 -5.43072820e-01 -4.60412055e-01
4.26415633e-03 9.02799815e-02 -1.59309238e-01 -5.17667904e-02
-9.68524277e-01 8.66442144e-01 2.50359960e-02 5.75181186e-01
-5.80013931e-01 -3.26822907e-01 3.66407186e-01 3.88596237e-01
6.28721654e-01 9.71356571e-01 -6.46760464e-01 -9.16246891e-01
-4.96229082e-01 1.62858739e-01 1.43626201e+00 1.01225817e+00
1.04341054e+00 -6.10616624e-01 7.76318908e-02 1.04095554e+00
2.34898195e-01 1.04548432e-01 7.83376336e-01 -8.79269302e-01
4.22401935e-01 1.14111257e+00 2.70960510e-01 -8.11528265e-01
-4.57083583e-01 2.32222024e-02 -3.12249005e-01 -4.53558832e-01
1.65645897e-01 -4.70417559e-01 -5.73234499e-01 1.61545491e+00
2.69347638e-01 1.31666675e-01 3.14258486e-01 6.80322707e-01
7.54086554e-01 8.01976740e-01 -7.34404773e-02 -4.03718442e-01
1.15584946e+00 -1.16061831e+00 -4.94220704e-01 -7.21922994e-01
1.01316726e+00 -6.04627132e-01 1.58176327e+00 3.68384868e-01
-5.49170256e-01 -4.37641054e-01 -8.91483784e-01 3.26243006e-02
-2.81795323e-01 -1.20960374e-03 7.23138750e-01 6.81522310e-01
-7.60488331e-01 1.26198441e-01 -5.68972290e-01 -7.72988498e-01
-1.16090380e-01 1.97606727e-01 -1.86861098e-01 -1.83771864e-01
-1.08704066e+00 4.99298602e-01 3.38969767e-01 -4.10045803e-01
-6.16030872e-01 -8.31934035e-01 -8.50593746e-01 6.47506416e-02
6.90549791e-01 -3.05465043e-01 1.84297085e+00 -7.55173743e-01
-1.78140128e+00 4.83647883e-01 -4.41891283e-01 -3.69687468e-01
1.28724426e-01 -3.72479200e-01 -2.06051663e-01 -2.73218572e-01
2.04729572e-01 4.30563062e-01 4.73730952e-01 -1.24492335e+00
-9.40420985e-01 -2.36221731e-01 4.71622229e-01 2.81523138e-01
-6.05633557e-01 -2.20619112e-01 -4.25067484e-01 -2.46072710e-01
-1.93853870e-01 -1.09102082e+00 -2.56950557e-01 -8.63835931e-01
-3.05149741e-02 -8.02834690e-01 9.59612966e-01 -5.39808750e-01
1.42027736e+00 -2.16357064e+00 5.77110201e-02 4.91091795e-02
3.80588770e-01 1.62606925e-01 -1.32750764e-01 7.16612518e-01
2.74596453e-01 6.42889440e-02 -3.48507673e-01 -5.59409678e-01
6.47403374e-02 3.15289125e-02 -1.92944795e-01 -2.08168864e-01
-1.17125742e-01 7.67833233e-01 -1.28779399e+00 -1.25020519e-01
8.93967897e-02 -8.97376314e-02 -8.09362173e-01 7.07453549e-01
-6.05189383e-01 3.53554964e-01 -4.69671428e-01 3.68353665e-01
1.04855426e-01 -4.67177421e-01 3.50252390e-01 2.48494204e-02
-8.19563642e-02 6.39234185e-01 -7.90867925e-01 1.95829403e+00
-9.99251723e-01 5.79055786e-01 -3.76110494e-01 -8.15286279e-01
9.58613932e-01 3.99149865e-01 3.97823662e-01 -3.69160354e-01
-4.40268824e-03 5.37627004e-02 1.18962996e-01 -7.11910248e-01
5.35068572e-01 5.01516536e-02 -5.56796551e-01 1.19781566e+00
1.20168559e-01 7.83766620e-03 3.37901682e-01 4.19280589e-01
1.47875750e+00 -6.00573003e-01 6.40519321e-01 -2.59737402e-01
5.42761087e-01 1.40205130e-01 4.11500007e-01 8.29732120e-01
-2.31877610e-01 2.77410626e-01 3.78005028e-01 -6.03867352e-01
-5.68889201e-01 -4.76357341e-01 4.03966576e-01 1.49165392e+00
-1.17041923e-01 -7.48162150e-01 -7.83195972e-01 -1.14188182e+00
-3.67793173e-01 1.18274808e+00 -5.99180639e-01 -6.12590797e-02
-3.57092738e-01 -3.28224063e-01 2.06564233e-01 2.25815669e-01
5.64242303e-01 -1.13944137e+00 -6.75897479e-01 4.33243662e-01
-6.25649035e-01 -1.47132945e+00 -6.59491658e-01 -4.27600779e-02
-3.90526265e-01 -9.36321855e-01 -2.33982086e-01 -6.57100558e-01
6.07387006e-01 3.64761263e-01 1.10974681e+00 7.22042546e-02
2.24920847e-02 5.11705577e-01 -8.45407009e-01 -2.88934708e-01
-7.38939285e-01 5.31050086e-01 3.50756645e-01 2.45165080e-01
9.43474054e-01 -6.09446883e-01 -2.38825008e-01 3.92864555e-01
-7.06543207e-01 1.55214697e-01 4.88322377e-01 6.18236184e-01
-2.73439735e-01 2.16021895e-01 7.66547561e-01 -1.26324773e+00
1.15376759e+00 -6.75707936e-01 -3.20807278e-01 2.60787517e-01
-7.04709768e-01 8.12552050e-02 7.69785345e-01 -3.92904043e-01
-1.30448818e+00 1.51022375e-01 5.07481806e-02 1.99861065e-01
-5.89818180e-01 8.70152950e-01 -1.06960172e-02 2.73336977e-01
8.16964090e-01 2.26129994e-01 -2.36415103e-01 -4.15809810e-01
5.83270490e-01 1.32407105e+00 2.27720261e-01 -5.81372976e-01
4.53631192e-01 1.23420589e-01 -8.51811349e-01 -9.87305105e-01
-9.08956766e-01 -1.09673393e+00 -7.50314653e-01 -2.69951195e-01
4.97380048e-01 -6.49740338e-01 -7.00589299e-01 1.93564832e-01
-1.26850283e+00 -4.72319067e-01 5.34460545e-02 3.25477958e-01
-4.10582423e-01 3.18893135e-01 -4.42334384e-01 -8.91298771e-01
-5.64540267e-01 -1.15883505e+00 7.95547664e-01 3.35319132e-01
-9.68510032e-01 -9.00566578e-01 3.89496058e-01 8.49474788e-01
4.16771442e-01 -3.25755507e-01 8.74276876e-01 -1.43341291e+00
-1.58188954e-01 -3.16688538e-01 -1.42095424e-02 1.78251326e-01
6.66084886e-01 -3.10848027e-01 -1.01299918e+00 -1.44049168e-01
1.66126251e-01 -4.57774907e-01 2.81980127e-01 -3.03056300e-01
5.76598048e-01 -7.05492616e-01 -3.80149424e-01 -3.54929477e-01
8.98175955e-01 6.42938137e-01 1.70817569e-01 -2.01036390e-02
5.59222043e-01 8.84250641e-01 7.19590783e-01 6.36580408e-01
7.73161352e-01 5.61194956e-01 -1.35910630e-01 3.35645646e-01
3.23168665e-01 -1.70794770e-01 2.68779993e-01 1.12149811e+00
2.01011017e-01 -1.80283964e-01 -1.33208930e+00 8.21129024e-01
-2.18167090e+00 -7.43800223e-01 5.64440787e-01 1.91171145e+00
1.03576326e+00 3.20170522e-01 3.40222210e-01 -1.13535300e-01
4.77313131e-01 -1.51102571e-02 -4.96555328e-01 -3.20227116e-01
8.12661529e-01 -1.85307965e-01 -1.00227617e-01 6.17214680e-01
-9.64625537e-01 1.12391376e+00 5.73236084e+00 3.86698157e-01
-9.75397706e-01 1.43386945e-01 5.95200598e-01 2.39260316e-01
-3.24409604e-01 -1.70508772e-02 -6.22747242e-01 4.49080348e-01
9.78978336e-01 -3.65184963e-01 6.21115565e-01 1.25177252e+00
1.94245219e-01 -2.11236954e-01 -1.50392020e+00 9.19650793e-01
3.72427225e-01 -1.30278814e+00 -2.22651646e-01 -1.11067131e-01
7.25638568e-01 1.82152852e-01 -2.44545341e-01 7.33262241e-01
5.53651333e-01 -6.39069021e-01 -3.68377380e-02 9.93949026e-02
1.76075667e-01 -6.80565059e-01 8.18014562e-01 9.15526330e-01
-9.28071558e-01 -9.65755731e-02 1.42913327e-01 -4.89316642e-01
1.72515586e-01 4.94584203e-01 -1.70189166e+00 1.74059868e-01
5.77278376e-01 7.20437765e-01 -3.49677473e-01 5.15980601e-01
-1.45787627e-01 8.52037489e-01 -3.79229873e-01 -5.03694952e-01
4.85202134e-01 -6.27074530e-03 5.21581411e-01 1.28045964e+00
-2.03049779e-01 5.19818246e-01 3.30251038e-01 7.58686721e-01
-3.49670112e-01 2.35117242e-01 -7.87318051e-01 -3.13939333e-01
5.79694927e-01 1.40868413e+00 -5.63818336e-01 -3.34682882e-01
-5.45736194e-01 9.97912705e-01 3.98458511e-01 2.01557860e-01
-3.54636937e-01 -3.95534068e-01 6.22192383e-01 -1.32596493e-01
-8.38293284e-02 -2.14722484e-01 -1.08597115e-01 -1.14925945e+00
1.02594033e-01 -1.11600721e+00 3.54260951e-01 -2.69454509e-01
-1.38804770e+00 7.99418628e-01 3.16130705e-02 -1.09664381e+00
-8.97172749e-01 -2.47323394e-01 -8.23198557e-01 4.08348471e-01
-1.14190578e+00 -8.37651789e-01 -3.91040802e-01 1.92801192e-01
1.29415822e+00 -3.77951056e-01 8.72419953e-01 1.50264680e-01
-4.80960190e-01 4.59311575e-01 -2.45884359e-01 1.97806239e-01
6.54977679e-01 -1.07920563e+00 8.75786006e-01 8.58938038e-01
2.24042997e-01 1.01800787e+00 8.13632131e-01 -6.54341340e-01
-1.48266137e+00 -8.41571867e-01 1.06189322e+00 -6.25954568e-01
7.18325555e-01 -5.60881615e-01 -9.18557465e-01 9.29484546e-01
3.53412718e-01 -5.25891900e-01 1.15495598e+00 6.44083798e-01
-3.21750551e-01 2.79256195e-01 -1.09738803e+00 7.16905832e-01
9.28667665e-01 -5.59934735e-01 -1.04252267e+00 3.94353658e-01
1.00184667e+00 -2.51814395e-01 -6.95668697e-01 3.44642222e-01
5.58200777e-01 -7.34823585e-01 4.38495606e-01 -6.88697696e-01
3.21201771e-01 -8.13655779e-02 -1.95971608e-01 -1.48086584e+00
-1.46586448e-01 -1.08846104e+00 3.13142464e-02 1.34499526e+00
7.89117396e-01 -4.49367523e-01 8.47767234e-01 1.23653531e+00
1.32789640e-02 -5.65494776e-01 -5.22651911e-01 -5.20399570e-01
-4.86838430e-01 -7.22363114e-01 4.95150954e-01 1.10127759e+00
9.63504136e-01 1.10436678e+00 -4.03603107e-01 4.90465648e-02
4.03651118e-01 1.93147272e-01 1.25376511e+00 -1.23714197e+00
-1.74496979e-01 -2.78800666e-01 -1.31256104e-01 -1.35573566e+00
3.43314141e-01 -6.04247987e-01 4.01806116e-01 -1.43902242e+00
1.60016105e-01 -4.41675752e-01 2.05241074e-03 5.80210090e-01
-2.81286448e-01 -2.52817690e-01 7.26015642e-02 3.10058922e-01
-9.84473944e-01 4.18891668e-01 4.65173513e-01 -3.45689327e-01
-8.57560515e-01 2.79703647e-01 -8.32372189e-01 9.83271241e-01
8.87692571e-01 -3.97068977e-01 -7.47687280e-01 -1.96669698e-01
1.98673885e-02 -2.16153711e-01 -3.68166476e-01 -9.38769519e-01
5.01053512e-01 -3.10386658e-01 -4.88471895e-01 -5.14936447e-01
2.79410422e-01 -8.99857461e-01 -3.30349863e-01 1.35413632e-01
-6.28336966e-01 -2.57459015e-01 -3.75272916e-03 3.57850939e-01
-1.35317340e-01 -3.70152265e-01 3.57538462e-01 -1.32186025e-01
-1.30829751e+00 2.98410989e-02 -5.99762440e-01 1.90022215e-01
1.05557263e+00 1.92871258e-01 -2.68900007e-01 -8.26140225e-01
-2.05033675e-01 3.72048795e-01 2.82889932e-01 7.58692682e-01
5.62237144e-01 -9.06967819e-01 -5.23552597e-01 9.10767093e-02
9.45517004e-01 -1.26812026e-01 5.13796136e-02 4.26706553e-01
-1.25580832e-01 5.66548884e-01 1.61588162e-01 -5.62269390e-01
-1.25203502e+00 3.77483040e-01 -1.56613439e-01 -2.64090776e-01
-3.47614288e-01 7.57646143e-01 1.06900960e-01 -8.36155117e-01
3.49653751e-01 -3.10553551e-01 -3.46896499e-01 -6.08434789e-02
7.56195545e-01 2.97219336e-01 -1.52968522e-03 -3.40402573e-01
-4.73184764e-01 8.51843804e-02 -5.10977268e-01 -2.66266853e-01
1.15260494e+00 -3.29399854e-01 5.41526563e-02 8.01229060e-01
1.37290323e+00 -1.90873310e-01 -8.46830547e-01 -7.56962180e-01
4.34605569e-01 -2.63277054e-01 -2.59705037e-01 -8.19546402e-01
-4.09120262e-01 3.88927221e-01 1.43086553e-01 5.45920968e-01
8.21984589e-01 4.08055097e-01 1.09249735e+00 1.09660470e+00
5.17484725e-01 -1.15016150e+00 2.90635109e-01 6.71552896e-01
8.20907235e-01 -1.72892153e+00 -3.06570858e-01 -3.40580195e-01
-1.07356477e+00 8.53675187e-01 8.11715424e-01 2.89793581e-01
6.74197257e-01 7.33349025e-02 4.53509778e-01 -1.55824274e-01
-1.05168808e+00 -1.02615692e-01 2.95659035e-01 3.57167870e-01
5.69391847e-01 5.11148907e-02 -2.68394232e-01 7.52151430e-01
-2.09249586e-01 1.20961722e-02 5.87986290e-01 1.15216684e+00
-5.18853664e-01 -1.16467547e+00 2.61641473e-01 3.87686640e-01
8.05794168e-03 -1.55218989e-01 -6.74541116e-01 4.27326292e-01
-2.88707405e-01 1.44361567e+00 -1.60837844e-01 -7.51894772e-01
2.89162308e-01 3.23687017e-01 -1.91205189e-01 -1.08475316e+00
-5.57859004e-01 -2.13863999e-01 3.10626477e-01 -5.01494586e-01
-6.10265672e-01 -7.57417738e-01 -1.20928562e+00 1.81066878e-02
-1.82869449e-01 5.59249938e-01 6.17354751e-01 1.36843717e+00
5.78386307e-01 7.11808503e-02 1.03918755e+00 -4.79101360e-01
-2.51558304e-01 -1.22949207e+00 2.88188048e-02 5.64226151e-01
1.85273707e-01 -4.36596334e-01 -4.24032509e-01 3.48235995e-01]
|
[12.488569259643555, 7.5520339012146]
|
9d4ffa26-2a83-4a0f-b60b-dd37b3e76a50
|
spae-semantic-pyramid-autoencoder-for
|
2306.17842
| null |
https://arxiv.org/abs/2306.17842v2
|
https://arxiv.org/pdf/2306.17842v2.pdf
|
SPAE: Semantic Pyramid AutoEncoder for Multimodal Generation with Frozen LLMs
|
In this work, we introduce Semantic Pyramid AutoEncoder (SPAE) for enabling frozen LLMs to perform both understanding and generation tasks involving non-linguistic modalities such as images or videos. SPAE converts between raw pixels and interpretable lexical tokens (or words) extracted from the LLM's vocabulary. The resulting tokens capture both the semantic meaning and the fine-grained details needed for visual reconstruction, effectively translating the visual content into a language comprehensible to the LLM, and empowering it to perform a wide array of multimodal tasks. Our approach is validated through in-context learning experiments with frozen PaLM 2 and GPT 3.5 on a diverse set of image understanding and generation tasks. Our method marks the first successful attempt to enable a frozen LLM to generate image content while surpassing state-of-the-art performance in image understanding tasks, under the same setting, by over 25%.
|
['Lu Jiang', 'Alexander G. Hauptmann', 'Kevin Murphy', 'Ming-Hsuan Yang', 'Yonatan Bisk', 'Irfan Essa', 'David A. Ross', 'Yanping Huang', 'Wolfgang Macherey', 'Vivek Kumar', 'Zhiruo Wang', 'Yong Cheng', 'Lijun Yu']
|
2023-06-30
| null | null | null | null |
['multimodal-generation']
|
['natural-language-processing']
|
[ 7.42054164e-01 5.05807579e-01 3.41574065e-02 -2.36047223e-01
-8.99081647e-01 -6.70258582e-01 1.01048017e+00 -1.40399888e-01
-2.15360835e-01 5.69649220e-01 4.02893186e-01 -3.13174576e-01
4.06930000e-01 -9.43743229e-01 -1.16917253e+00 -4.48271722e-01
2.93184191e-01 2.78107256e-01 -2.88079321e-01 -2.23466560e-01
2.64977571e-02 1.54663861e-01 -1.74156082e+00 9.24234390e-01
4.80432272e-01 1.18332481e+00 6.37578368e-01 6.86977267e-01
-4.08260077e-01 1.16037393e+00 -5.66014051e-01 -6.52770579e-01
-2.16031969e-02 -2.17633620e-01 -1.07584906e+00 3.39176089e-01
7.30808198e-01 -4.51296389e-01 -2.63493896e-01 8.07986975e-01
1.27759784e-01 -1.18956128e-02 7.06676781e-01 -1.21925986e+00
-1.11158085e+00 8.59859228e-01 -2.43249208e-01 -4.11475033e-01
7.52565086e-01 4.27068532e-01 1.00805235e+00 -9.05595899e-01
8.85812521e-01 1.49422061e+00 4.75014418e-01 7.94284284e-01
-1.16000926e+00 -5.97216785e-01 -9.64216143e-02 1.00139417e-01
-1.10181582e+00 -6.91923320e-01 5.40458918e-01 -2.36384794e-01
1.17557704e+00 1.22248717e-01 7.30948329e-01 1.58006322e+00
2.03103740e-02 8.22698116e-01 1.27928090e+00 -5.66044152e-01
1.90847307e-01 2.08304301e-01 -4.67779487e-01 8.24679613e-01
-1.75176248e-01 -8.89906939e-03 -1.13394749e+00 2.24836335e-01
8.89591336e-01 -4.29294765e-01 -2.94789284e-01 5.13281897e-02
-1.46784520e+00 7.64212728e-01 4.23233360e-01 1.75590679e-01
-7.23514676e-01 5.50068915e-01 2.29131162e-01 1.62871942e-01
1.52974740e-01 4.24775571e-01 -2.95216680e-01 -1.96404874e-01
-8.32376420e-01 1.49892699e-02 8.67550671e-01 8.97169769e-01
6.65548563e-01 2.17270523e-01 -2.49672115e-01 4.07638192e-01
4.12976950e-01 8.20729852e-01 2.79440194e-01 -1.30055511e+00
4.96741176e-01 5.74766576e-01 2.96101756e-02 -8.60761046e-01
2.15731695e-01 -6.26414875e-03 -7.11922705e-01 3.72355819e-01
3.15289676e-01 6.72287047e-02 -1.20788670e+00 1.90937042e+00
-2.67997272e-02 1.00306002e-02 4.95767921e-01 8.01149249e-01
1.13466299e+00 1.09473705e+00 5.88485777e-01 2.78687000e-01
1.72978950e+00 -7.18566656e-01 -5.09794354e-01 -5.00041127e-01
3.01307384e-02 -6.73437178e-01 1.33833647e+00 4.03338313e-01
-1.26029086e+00 -7.89491713e-01 -8.61714900e-01 -2.85250932e-01
-4.42476392e-01 1.41053557e-01 5.99040329e-01 5.10775328e-01
-1.41504002e+00 3.79887909e-01 -4.80493754e-01 -4.73215193e-01
6.75588787e-01 1.69358075e-01 -8.16540420e-01 -2.55735844e-01
-1.22415483e+00 8.63885701e-01 6.70224190e-01 -2.56182045e-01
-1.23077595e+00 -9.29767489e-01 -1.23032618e+00 1.00750558e-01
1.68441206e-01 -1.02566409e+00 1.25542188e+00 -1.34651017e+00
-1.50967705e+00 1.23853278e+00 -2.01996386e-01 -7.11206257e-01
2.77281582e-01 -1.85190618e-01 -1.01699106e-01 5.35487592e-01
1.05766334e-01 1.70431292e+00 1.32851017e+00 -1.65197945e+00
-4.46495384e-01 -1.35477096e-01 3.93352598e-01 4.46722835e-01
-2.68348306e-01 -3.67084205e-01 -5.28529465e-01 -5.64395368e-01
-5.15617430e-01 -6.61930084e-01 1.65852964e-01 5.65604307e-02
-4.18066084e-01 -8.12200829e-02 9.71485496e-01 -1.05819952e+00
5.21629930e-01 -1.96488369e+00 4.81516868e-01 -4.03725021e-02
2.81381339e-01 2.15619504e-01 -5.45397341e-01 5.49344540e-01
-9.51536074e-02 2.45763302e-01 -2.55613714e-01 -7.73052871e-01
2.19661877e-01 4.09884542e-01 -7.08889008e-01 -1.73378214e-01
4.81064707e-01 1.28413427e+00 -8.71827960e-01 -4.61965472e-01
7.65915334e-01 8.47811460e-01 -2.70388782e-01 2.77503878e-01
-7.16723621e-01 5.32993555e-01 -3.73481005e-01 5.32849252e-01
3.85331273e-01 -3.69443119e-01 2.10818753e-01 -6.23916864e-01
1.88161269e-01 -9.94510055e-02 -6.90179169e-01 2.04200029e+00
-1.01731181e+00 8.83944571e-01 -4.80827801e-02 -6.87571287e-01
5.61495721e-01 3.63723278e-01 1.17376447e-01 -8.55488479e-01
4.59001772e-02 -1.37984589e-01 -5.99100590e-01 -6.05994821e-01
6.39655888e-01 -3.55565250e-01 -2.11264625e-01 5.58817089e-01
3.94262344e-01 -4.14231122e-01 6.38298169e-02 3.70508194e-01
6.79173768e-01 4.80683774e-01 1.72622472e-01 -5.68707241e-03
5.93320847e-01 -8.15771669e-02 -3.00324559e-01 8.90826464e-01
1.95146739e-01 5.16031682e-01 2.98978239e-01 -3.88588071e-01
-1.22274220e+00 -1.34166849e+00 4.36784297e-01 1.04932356e+00
-4.68507558e-02 -3.76700819e-01 -1.02693808e+00 -3.75508666e-01
-1.18914261e-01 1.09694195e+00 -7.92232156e-01 -1.28308386e-01
-1.30014569e-01 -1.02835551e-01 6.98406935e-01 5.47899067e-01
8.30280304e-01 -1.68523121e+00 -8.37194920e-01 -1.88217890e-02
-6.05297446e-01 -1.71382773e+00 -1.14366896e-01 -2.80092895e-01
-4.84103888e-01 -9.32710886e-01 -5.92862606e-01 -7.17845082e-01
5.93164921e-01 1.42565509e-03 1.60816407e+00 -6.48134127e-02
-3.48781317e-01 8.63040030e-01 -4.40606415e-01 -3.36760223e-01
-8.47222865e-01 -8.33073035e-02 -4.17775661e-01 1.27854124e-01
5.33392169e-02 -4.14992541e-01 -4.32498127e-01 -2.34086409e-01
-1.35152662e+00 7.21526146e-01 7.51390755e-01 7.43174672e-01
5.68611085e-01 -1.17994204e-01 2.29895562e-01 -4.69751775e-01
5.72814107e-01 -1.58116758e-01 -2.19191164e-01 4.33823764e-01
-7.32172728e-02 2.89994717e-01 6.10291898e-01 -3.72421354e-01
-1.39001119e+00 7.47418702e-02 -1.09224007e-01 -3.86216640e-01
-3.91830415e-01 3.36755484e-01 -1.84154287e-01 -1.94713816e-01
4.91534799e-01 6.37641966e-01 5.03412187e-02 -1.18920304e-01
1.01466954e+00 6.06911778e-01 8.13254833e-01 -6.83063030e-01
6.85497701e-01 7.69284189e-01 -7.87217095e-02 -8.62594545e-01
-8.31244349e-01 5.08372858e-02 -4.34181988e-01 -3.73727083e-01
1.23532236e+00 -1.06738245e+00 -8.62079322e-01 5.79782605e-01
-1.29768980e+00 -5.95459163e-01 -4.73706782e-01 -1.34050623e-01
-8.53710473e-01 3.82370651e-01 -4.38950390e-01 -6.19693875e-01
-5.82891822e-01 -1.27905846e+00 1.65513313e+00 3.13238382e-01
-2.24511683e-01 -1.11291623e+00 -3.30880880e-01 9.43364918e-01
3.76073986e-01 2.49938115e-01 9.46517885e-01 -2.12162863e-02
-7.47275531e-01 4.14820425e-02 -6.38080120e-01 5.29198170e-01
-1.56670779e-01 -2.50474304e-01 -1.20981109e+00 7.09442934e-03
-3.29474151e-01 -9.32507336e-01 9.35993075e-01 2.57992268e-01
1.27226353e+00 -1.85975015e-01 2.41328198e-02 4.80199397e-01
1.47894418e+00 -1.04384400e-01 1.02030313e+00 1.86318427e-01
6.64156437e-01 5.48672140e-01 2.11808309e-01 3.92828554e-01
5.86821437e-01 5.03042102e-01 6.03088319e-01 -1.87811747e-01
-5.73329508e-01 -6.61420822e-01 5.76170623e-01 2.55673379e-01
-5.39002717e-02 -5.25471449e-01 -9.14182603e-01 5.60114503e-01
-1.59061992e+00 -8.76733124e-01 2.55858898e-01 1.69856954e+00
9.56722379e-01 -2.45235056e-01 -3.85239303e-01 -2.96675205e-01
3.54251951e-01 3.96910965e-01 -4.28090066e-01 -4.51596588e-01
-2.38871530e-01 5.38202167e-01 2.67360181e-01 5.44982970e-01
-9.96068537e-01 1.42633069e+00 6.25753593e+00 1.02258670e+00
-9.89850402e-01 -3.27089727e-02 7.02049553e-01 3.76767069e-01
-4.71948475e-01 -5.32282218e-02 -3.21688712e-01 1.32145464e-01
9.71109331e-01 2.10942358e-01 7.90105999e-01 5.02213240e-01
4.15804870e-02 -2.36167789e-01 -9.95349169e-01 1.17306578e+00
3.62030506e-01 -1.57923901e+00 7.78865576e-01 -2.20559552e-01
6.48338675e-01 -2.64640659e-01 3.60947967e-01 2.67554522e-01
2.77594715e-01 -1.56650627e+00 1.04735291e+00 7.22751856e-01
1.22287393e+00 -3.90000701e-01 5.33616722e-01 6.26291260e-02
-1.14911020e+00 2.02003345e-02 1.59268118e-02 -3.69236730e-02
1.92099214e-01 2.55554885e-01 -8.68467450e-01 5.30648887e-01
6.33816779e-01 6.95030153e-01 -4.40544099e-01 4.65395525e-02
-4.97566581e-01 5.23968458e-01 -1.68627381e-01 2.10914612e-01
4.84105319e-01 1.58801645e-01 4.37037587e-01 1.14433897e+00
2.51555592e-01 -8.27666279e-03 -1.17233559e-01 1.09117222e+00
-4.03985500e-01 -1.23514429e-01 -6.41715944e-01 -5.27434826e-01
3.18478882e-01 1.20324898e+00 -3.73666763e-01 -6.05777264e-01
-2.28468582e-01 1.36981773e+00 1.42504767e-01 5.50035417e-01
-7.52448559e-01 -3.51896547e-02 4.86851990e-01 -1.52909756e-01
2.10644320e-01 -4.35364582e-02 -1.52873904e-01 -1.30219877e+00
-1.18114889e-01 -1.09238517e+00 9.93978530e-02 -1.65305638e+00
-1.14726043e+00 6.66908205e-01 1.40038371e-01 -6.42879128e-01
-4.95705664e-01 -9.23330367e-01 -3.86598468e-01 8.10387731e-01
-1.49060738e+00 -2.02486849e+00 -6.24467909e-01 8.06558967e-01
7.81836927e-01 -2.35853389e-01 1.14589584e+00 -2.35557869e-01
3.55570801e-02 1.74436778e-01 -4.86659765e-01 8.94372985e-02
4.65499192e-01 -1.24785221e+00 6.09056234e-01 5.57048619e-01
3.55284095e-01 3.85875404e-01 7.51252353e-01 -4.58739936e-01
-1.68758905e+00 -8.92804325e-01 6.61474049e-01 -4.71507877e-01
6.71633840e-01 -4.10036355e-01 -4.03649956e-01 6.95763886e-01
6.91482961e-01 -4.84492838e-01 6.82852745e-01 -3.58611763e-01
-6.85078859e-01 1.56251892e-01 -1.06714189e+00 7.63158977e-01
1.02053714e+00 -8.70591938e-01 -7.13835061e-01 2.19409600e-01
7.80669510e-01 -3.46189290e-01 -9.98262107e-01 8.73958170e-02
7.20154643e-01 -8.66100848e-01 1.34513712e+00 -6.22742474e-01
1.11773109e+00 -6.52401373e-02 -5.63899994e-01 -1.19664705e+00
1.22110792e-01 -5.67267120e-01 -1.26768708e-01 1.06613159e+00
2.17835367e-01 -1.64905474e-01 6.38301432e-01 5.29521286e-01
8.55402723e-02 -4.29534703e-01 -7.16191113e-01 -1.65156871e-01
-8.46278444e-02 -8.45936835e-01 5.59526145e-01 6.23115361e-01
-3.79993141e-01 2.09350973e-01 -4.57138479e-01 -1.42006591e-01
9.04992998e-01 1.65056229e-01 7.64035583e-01 -7.20587850e-01
-4.04249638e-01 -3.00449997e-01 -3.09118718e-01 -9.42715883e-01
5.56537628e-01 -9.05257463e-01 -5.08721285e-02 -1.63278198e+00
2.43906721e-01 -1.17920833e-02 8.84515047e-02 8.39296043e-01
8.34600553e-02 6.09507263e-01 6.33260489e-01 -4.73732464e-02
-5.70354402e-01 4.58298028e-01 1.38801229e+00 -4.28066760e-01
1.64697737e-01 -6.51280820e-01 -9.87691760e-01 5.93248010e-01
6.39723539e-01 6.53147101e-02 -5.09855628e-01 -6.73653245e-01
2.31710389e-01 1.63858384e-01 1.02365601e+00 -8.34953189e-01
-1.76252604e-01 -1.16482280e-01 6.76605642e-01 -2.54596263e-01
6.13033175e-01 -7.24754035e-01 2.40068525e-01 2.89092124e-01
-5.22586107e-01 -3.03930163e-01 5.10526717e-01 4.74758118e-01
-2.94282258e-01 6.89702621e-03 5.13794243e-01 -3.99540693e-01
-1.38776171e+00 1.36676952e-01 -3.92759025e-01 -1.73796877e-01
8.20152223e-01 -3.49186957e-01 -2.90810317e-01 -9.02773798e-01
-9.53640044e-01 -5.97228818e-02 5.85941195e-01 5.62011540e-01
1.00562632e+00 -1.19632590e+00 -8.00515711e-01 2.26921692e-01
2.26631001e-01 -2.70486623e-01 6.04151607e-01 1.89105108e-01
-6.39965773e-01 5.07326484e-01 -5.26568651e-01 -4.71257061e-01
-1.16259658e+00 1.98578089e-01 2.19642237e-01 -3.55566621e-01
-6.37997210e-01 8.69435191e-01 4.07901257e-01 -2.13716194e-01
-3.79694655e-04 -1.47071153e-01 7.73122311e-02 -7.19119683e-02
5.51677704e-01 -2.01253429e-01 -3.33740920e-01 -8.16695750e-01
-9.91756320e-02 6.05267107e-01 3.05186123e-01 -5.40438950e-01
1.18105531e+00 -2.24914402e-01 -5.60126901e-02 1.57793730e-01
1.09229922e+00 -3.87434721e-01 -1.42090464e+00 3.47615825e-03
-5.59165657e-01 -2.14490980e-01 2.38232583e-01 -1.16477466e+00
-9.41308141e-01 1.06208253e+00 5.02823651e-01 -5.49267158e-02
1.35200691e+00 3.38663757e-01 1.08062959e+00 4.40469414e-01
4.41840589e-01 -7.68924057e-01 3.75930667e-01 2.89000273e-01
1.14610851e+00 -1.33204353e+00 -3.08016926e-01 -2.63640046e-01
-1.01843476e+00 1.21060169e+00 2.75604486e-01 1.31076083e-01
8.59279633e-02 7.07500279e-02 7.03972280e-02 -2.31582969e-01
-7.95764923e-01 -1.30661368e-01 4.50580299e-01 1.02619505e+00
2.30105177e-01 2.22040802e-01 5.60974956e-01 4.06099856e-01
-2.13025287e-01 1.75057516e-01 3.17058384e-01 6.29001379e-01
-2.51698554e-01 -1.05035341e+00 -3.57256889e-01 1.45635232e-01
-2.64632612e-01 -3.16077888e-01 -5.42409778e-01 9.53462303e-01
2.34406903e-01 8.73097301e-01 1.79181963e-01 -2.74530470e-01
-4.86631431e-02 2.18669266e-01 7.99414992e-01 -3.56423169e-01
-4.02174264e-01 -3.13005388e-01 1.42823815e-01 -7.10061967e-01
-6.66696012e-01 -4.72023129e-01 -1.17760956e+00 -1.56932548e-01
3.99050623e-01 -3.93095016e-01 8.25033605e-01 1.05966103e+00
3.17919880e-01 5.00757158e-01 7.02591985e-02 -1.00876105e+00
-1.18038952e-01 -8.20408285e-01 -1.72645196e-01 8.42430711e-01
7.44026452e-02 -2.56484956e-01 2.80083921e-02 8.07372868e-01]
|
[10.991865158081055, 1.2394617795944214]
|
8b0ec59b-0eb1-4eb1-83e6-b2f86b2fb95b
|
tiny-always-on-and-fragile-bias-propagation
|
2201.07677
| null |
https://arxiv.org/abs/2201.07677v4
|
https://arxiv.org/pdf/2201.07677v4.pdf
|
Tiny, always-on and fragile: Bias propagation through design choices in on-device machine learning workflows
|
Billions of distributed, heterogeneous and resource constrained IoT devices deploy on-device machine learning (ML) for private, fast and offline inference on personal data. On-device ML is highly context dependent, and sensitive to user, usage, hardware and environment attributes. This sensitivity and the propensity towards bias in ML makes it important to study bias in on-device settings. Our study is one of the first investigations of bias in this emerging domain, and lays important foundations for building fairer on-device ML. We apply a software engineering lens, investigating the propagation of bias through design choices in on-device ML workflows. We first identify reliability bias as a source of unfairness and propose a measure to quantify it. We then conduct empirical experiments for a keyword spotting task to show how complex and interacting technical design choices amplify and propagate reliability bias. Our results validate that design choices made during model training, like the sample rate and input feature type, and choices made to optimize models, like light-weight architectures, the pruning learning rate and pruning sparsity, can result in disparate predictive performance across male and female groups. Based on our findings we suggest low effort strategies for engineers to mitigate bias in on-device ML.
|
['Akhil Mathur', 'Aaron Yi Ding', 'Fahim Kawsar', 'Wiebke Toussaint']
|
2022-01-19
| null | null | null | null |
['keyword-spotting']
|
['speech']
|
[ 2.64008254e-01 4.69156951e-02 -7.04938054e-01 -6.38395250e-01
-4.67806607e-01 -5.98053575e-01 2.34694883e-01 2.46356443e-01
-2.31905162e-01 3.30700755e-01 3.63600463e-01 -8.68676543e-01
-2.26528525e-01 -5.94270289e-01 -5.90927064e-01 -7.34989643e-02
2.65708804e-01 1.87415347e-01 -5.06707668e-01 2.07105771e-01
3.04785579e-01 2.95500875e-01 -1.47832167e+00 2.51762688e-01
5.53357601e-01 1.02926528e+00 -1.53278410e-01 5.91061532e-01
2.04012498e-01 5.75545430e-01 -8.75059366e-01 -6.63583040e-01
2.14178756e-01 2.17852965e-01 -2.36165524e-01 -5.02694488e-01
5.25801003e-01 -3.56290072e-01 1.32875443e-01 6.81637526e-01
1.09360087e+00 -4.94123310e-01 4.12312359e-01 -1.40767479e+00
-5.97766161e-01 1.07890308e+00 -2.80723333e-01 2.79684693e-01
3.46609443e-01 5.17319262e-01 1.04694593e+00 -3.97766560e-01
2.83759892e-01 1.17547715e+00 1.00076091e+00 4.74846154e-01
-1.56104410e+00 -1.07183862e+00 1.73854977e-01 -4.53942031e-01
-1.32998800e+00 -8.13989043e-01 7.75079489e-01 -7.14441717e-01
9.71042037e-01 5.76794803e-01 6.03788912e-01 1.64572322e+00
4.18772697e-01 4.74622697e-01 1.04567611e+00 -3.84798616e-01
5.02153933e-01 7.26676643e-01 2.33566165e-01 4.84388769e-01
9.82904315e-01 1.94161221e-01 -7.68334270e-01 -8.22895586e-01
2.72698134e-01 1.96924612e-01 2.59492010e-01 -1.06495075e-01
-9.28114653e-01 5.18735170e-01 -1.31959125e-01 1.83370888e-01
-1.07005559e-01 3.57752055e-01 2.71410257e-01 2.46208742e-01
2.47995406e-01 9.93660390e-01 -1.07173145e+00 -6.04249299e-01
-1.04976749e+00 1.55592695e-01 1.11464846e+00 9.27114487e-01
3.11798602e-01 -2.34506771e-01 -5.96574187e-01 6.15323246e-01
4.01290178e-01 7.51716495e-01 1.28658518e-01 -8.01014483e-01
4.82865840e-01 6.14741445e-01 1.27117202e-01 -1.08348274e+00
-4.84599531e-01 -5.10896742e-01 -4.73611534e-01 -1.19494691e-01
1.99190900e-01 -5.10260463e-01 -4.37343240e-01 1.74153674e+00
-1.56922311e-01 -3.75276327e-01 -6.31583691e-01 4.12897736e-01
4.82136101e-01 -1.65653884e-01 5.94957173e-01 -5.61951147e-03
1.39242852e+00 -9.58529934e-02 -5.41121304e-01 -3.32951248e-01
7.99676061e-01 -7.05726981e-01 1.76579988e+00 5.14689326e-01
-9.69924450e-01 -3.45042646e-01 -1.12540329e+00 4.40603755e-02
-2.83597499e-01 3.80113214e-01 8.78447413e-01 1.59411490e+00
-4.69534785e-01 7.77551651e-01 -7.90562987e-01 -4.22401369e-01
9.17035580e-01 5.67883790e-01 4.76468205e-01 1.34490073e-01
-6.48478806e-01 7.93908238e-01 -3.91996056e-01 -3.73968601e-01
-6.31850839e-01 -1.23006761e+00 -2.48323306e-01 7.28442520e-02
2.39838868e-01 -9.35462892e-01 1.03924572e+00 -6.66381717e-01
-1.24525344e+00 6.64116204e-01 1.37435822e-02 -8.25859979e-02
2.58621544e-01 -2.25561008e-01 -7.81161427e-01 -8.75116348e-01
5.08429529e-03 1.04275972e-01 9.46210563e-01 -1.09537435e+00
-3.46043855e-01 -5.29357314e-01 -7.57162571e-02 -5.24447918e-01
-7.08029151e-01 9.47910249e-02 1.42321572e-01 -4.49085087e-01
-4.33879256e-01 -9.94800627e-01 -1.05564840e-01 -4.42235440e-01
-5.78402042e-01 -1.03755690e-01 5.62847555e-01 -6.49316132e-01
1.90750515e+00 -2.09940386e+00 -3.41433167e-01 4.35912579e-01
5.19557476e-01 -1.35610163e-01 3.07079762e-01 6.95450157e-02
3.17855835e-01 8.21500599e-01 4.43328679e-01 -3.33347410e-01
3.56511980e-01 -1.95082918e-01 -1.36288211e-01 2.81623542e-01
3.60477045e-02 9.24875915e-01 -4.64968175e-01 -4.06218171e-01
1.19955823e-01 3.48550081e-01 -9.14899528e-01 -1.65460501e-02
-1.07817389e-01 2.91905105e-01 -4.42059755e-01 1.20820796e+00
3.97417009e-01 -4.51041013e-01 3.08451712e-01 -4.53470916e-01
-9.04143527e-02 4.72211152e-01 -8.23828340e-01 1.32479596e+00
-9.50372636e-01 3.25517893e-01 1.44063860e-01 -3.22750807e-01
7.82140076e-01 -1.28930688e-01 3.36994141e-01 -6.88878238e-01
5.04977107e-01 2.12940350e-01 1.47864237e-01 -5.67479312e-01
4.72216159e-01 1.49002478e-01 -4.97954011e-01 5.52361369e-01
-3.16140145e-01 -1.33709311e-01 -4.25170392e-01 -4.47769091e-02
1.44140768e+00 -3.59662801e-01 1.92653000e-01 -5.90251148e-01
-2.13935345e-01 -5.45232534e-01 5.13848484e-01 1.02587783e+00
-2.44031399e-01 1.12528622e-01 4.72992122e-01 -4.37422097e-01
-9.38523054e-01 -8.35418820e-01 -2.56386608e-01 1.58450866e+00
-1.90979227e-01 -7.45073318e-01 -5.00060380e-01 -8.57884824e-01
6.64731920e-01 1.07127440e+00 -4.62802887e-01 -3.46584320e-01
-2.08504841e-01 -8.29863250e-01 7.06821799e-01 5.10189116e-01
8.60788859e-03 -5.40528059e-01 -9.18504477e-01 -1.55825481e-01
3.04587007e-01 -8.96223128e-01 -3.58335167e-01 2.92900890e-01
-9.84705269e-01 -8.46763849e-01 5.39425090e-02 -4.66059381e-03
6.04626417e-01 -2.96900541e-01 1.68939507e+00 1.26492649e-01
-5.61831892e-01 5.34414291e-01 1.09522469e-01 -1.05196369e+00
-2.86901444e-01 3.93675506e-01 1.87907502e-01 -3.69566470e-01
7.02030778e-01 -8.36034238e-01 -6.57966018e-01 5.43775558e-01
-3.39908749e-01 -3.18160057e-01 9.57802117e-01 3.46445501e-01
2.68222660e-01 4.00621211e-03 2.60065675e-01 -9.25950229e-01
1.08113849e+00 -6.54400110e-01 -4.20796901e-01 2.76256979e-01
-1.35497034e+00 1.30845293e-01 2.21704409e-01 -6.85770214e-01
-6.68617070e-01 -2.52040744e-01 2.62970299e-01 -9.40709487e-02
-9.34815779e-02 -4.82281409e-02 -5.27724683e-01 -5.22494130e-03
9.45329607e-01 -6.87920034e-01 -4.59736079e-01 -5.96787155e-01
2.10151926e-01 9.40216005e-01 -2.41359636e-01 -9.48256969e-01
3.87644112e-01 3.97637278e-01 -2.14878082e-01 -5.90854406e-01
-3.93675745e-01 2.24526256e-01 -1.51309505e-01 4.28498201e-02
5.46633065e-01 -8.33679557e-01 -1.18046904e+00 -1.48133725e-01
-6.31375194e-01 -5.15014172e-01 -3.60960305e-01 1.75811589e-01
-6.90955669e-02 -3.86204779e-01 -1.09924428e-01 -1.11476088e+00
-5.30369818e-01 -1.19188857e+00 1.22825384e+00 1.53606653e-01
-1.11529613e+00 -7.32068062e-01 -3.59287858e-01 4.58179504e-01
9.02103841e-01 4.12283652e-02 1.02433777e+00 -6.18736804e-01
-3.68269295e-01 -3.06325585e-01 -3.20868604e-02 1.64343849e-01
4.70526926e-02 3.14917825e-02 -1.02380300e+00 -2.11785257e-01
-8.67839456e-02 -6.36005774e-02 2.17421770e-01 4.93845195e-01
1.61384261e+00 -5.14413178e-01 -7.03263223e-01 6.09623849e-01
1.17887557e+00 -1.44761801e-01 1.43734574e-01 4.00578044e-02
8.64561856e-01 3.25175464e-01 2.83058375e-01 7.88856268e-01
3.31055969e-01 6.51071012e-01 7.69145191e-02 1.36127576e-01
1.18356161e-01 -5.25428474e-01 1.98171586e-01 3.30607146e-01
6.37633055e-02 -1.10855550e-01 -9.78850067e-01 2.65234530e-01
-1.43485749e+00 -4.64057744e-01 2.97721356e-01 2.44556499e+00
7.02926636e-01 4.89864200e-01 3.60703707e-01 1.02164738e-01
5.01777768e-01 -5.16313966e-03 -7.87248433e-01 -6.30042195e-01
7.07714558e-02 2.39952490e-01 1.12749445e+00 1.00767583e-01
-7.18293965e-01 3.92343640e-01 7.18105030e+00 4.63553935e-01
-1.09216690e+00 3.77192229e-01 1.13501918e+00 -8.81068587e-01
-6.99363887e-01 -1.09663673e-01 -9.97946084e-01 9.56396997e-01
1.21885586e+00 -1.33887669e-02 6.43432736e-01 1.27127814e+00
3.40369672e-01 7.94731751e-02 -1.69263875e+00 1.24492800e+00
-2.24863380e-01 -1.19630623e+00 -4.83939648e-01 4.54488724e-01
5.20488560e-01 -8.70199800e-02 4.02802765e-01 3.24360549e-01
5.43903291e-01 -1.30797863e+00 7.94219375e-01 3.25201392e-01
8.85282815e-01 -6.09425843e-01 6.03375554e-01 4.51063849e-02
-5.64121068e-01 -5.53086042e-01 1.08726084e-01 -4.61541951e-01
-1.10160284e-01 1.18771982e+00 -9.60043132e-01 -3.61205101e-01
1.03817225e+00 2.08350971e-01 -7.93277144e-01 2.78358459e-01
1.93801433e-01 1.02406800e+00 -4.39995557e-01 -4.80123729e-01
-8.55762362e-01 4.68066156e-01 3.20712239e-01 1.25153649e+00
2.24177212e-01 -2.66949236e-01 -1.77868143e-01 1.20909381e+00
-1.73423007e-01 -1.72186807e-01 -6.03712201e-01 -4.46287572e-01
1.00125718e+00 1.23554790e+00 -4.39025342e-01 4.15816344e-02
-2.18934223e-01 4.07172531e-01 -8.00946578e-02 1.78737611e-01
-6.91576064e-01 2.13215929e-02 1.04945707e+00 5.09174347e-01
-9.74070467e-03 -1.66620746e-01 -1.44290459e+00 -7.84615636e-01
2.29440182e-01 -1.16984940e+00 1.03876911e-01 -2.08523422e-01
-1.38281071e+00 -7.15924054e-02 1.23107936e-02 -6.75960720e-01
1.58702359e-01 -2.43559420e-01 -5.41565299e-01 6.52155101e-01
-8.19243908e-01 -1.11467445e+00 -2.95908660e-01 1.45472586e-01
8.05066153e-02 -3.09082776e-01 6.87373400e-01 3.56771410e-01
-7.92655528e-01 1.29159951e+00 -3.80486339e-01 -3.67041498e-01
6.45258963e-01 -8.49874496e-01 6.45034075e-01 3.95986766e-01
-1.27989426e-01 1.32650316e+00 8.08061898e-01 -8.94527018e-01
-1.87720168e+00 -7.84173489e-01 7.80621827e-01 -1.30557358e+00
4.14539248e-01 -8.78088713e-01 -1.64293170e-01 6.06374741e-01
-4.27853793e-01 -1.49041370e-01 1.18786871e+00 9.00482297e-01
-4.41739142e-01 -3.83339971e-01 -1.54452765e+00 6.13300085e-01
1.47419369e+00 -5.86765468e-01 -4.60930401e-03 1.52332097e-01
9.38438833e-01 -9.26290266e-03 -1.34638965e+00 4.80292350e-01
1.11528182e+00 -8.91488552e-01 9.46400046e-01 -5.24116635e-01
2.97864050e-01 2.70456642e-01 -4.40956265e-01 -7.29404867e-01
-4.52091068e-01 -1.03735077e+00 -4.38064247e-01 1.66021299e+00
8.69824708e-01 -5.21323860e-01 9.80929017e-01 1.52727568e+00
2.15737253e-01 -9.45333540e-01 -5.15059412e-01 -5.41684031e-01
-1.17646828e-01 -8.68346035e-01 1.07737458e+00 7.49165833e-01
-1.13070138e-01 3.32565874e-01 -2.87856370e-01 -8.13234132e-03
5.53390861e-01 -3.02677482e-01 8.48740518e-01 -1.39667416e+00
-5.53108633e-01 -4.43339527e-01 -9.33957174e-02 -4.87283140e-01
-1.23913109e-01 -5.67451715e-01 -4.65942204e-01 -8.61515701e-01
-2.67303530e-02 -1.13712609e+00 -3.83671373e-01 3.98974836e-01
1.54816667e-02 -1.85116678e-01 1.16140231e-01 2.25296944e-01
-4.39165384e-01 -4.47491445e-02 3.84043187e-01 -1.51599348e-01
-5.65789938e-01 2.41854802e-01 -1.42770362e+00 6.83452666e-01
8.53473008e-01 -5.47031105e-01 -3.14393520e-01 -4.06538367e-01
7.10962355e-01 -5.96853852e-01 3.06388050e-01 -9.51359689e-01
-6.59210086e-02 -2.01932684e-01 6.28819764e-01 -1.02287754e-01
-4.70200405e-02 -1.03684473e+00 1.30999461e-01 3.78328413e-01
-5.08163333e-01 6.51842430e-02 2.68456519e-01 4.51241970e-01
8.85750830e-01 3.11274916e-01 2.39772484e-01 1.55997515e-01
-6.91818027e-03 -9.96229704e-03 -4.16657701e-02 1.10213093e-01
7.12113142e-01 1.90917845e-03 -2.38534540e-01 -1.34401284e-02
-3.67638111e-01 -1.41528964e-01 7.22242892e-01 6.77268445e-01
1.37000214e-02 -1.02731049e+00 -1.86745808e-01 5.00079393e-01
1.89436868e-01 -5.80052078e-01 -6.69034570e-02 6.90078974e-01
1.50366947e-01 3.24307293e-01 1.25513569e-01 -5.00324428e-01
-1.33200276e+00 4.07701641e-01 -4.64965552e-02 -6.28532022e-02
-1.16050377e-01 1.01237631e+00 -5.12183905e-01 -3.24454695e-01
4.98423100e-01 -6.02398872e-01 5.54454505e-01 -1.10736847e-01
2.04499066e-01 9.40016270e-01 3.20197612e-01 2.00511336e-01
-6.48658931e-01 3.29128534e-01 2.58947611e-01 -3.85564677e-02
9.77144539e-01 -1.12243235e-01 7.01189712e-02 5.94781399e-01
1.00237656e+00 5.53562939e-01 -8.65216553e-01 2.52543658e-01
4.56830636e-02 -5.74019372e-01 2.58879215e-01 -1.17101514e+00
-7.98911989e-01 4.67789054e-01 9.63528454e-01 2.80663192e-01
9.19919968e-01 8.13636556e-02 6.94854856e-01 2.30642006e-01
6.41278923e-01 -1.40970254e+00 -1.49481833e-01 -3.48438531e-01
5.30128479e-01 -1.30211759e+00 4.91888970e-01 -2.77296484e-01
-4.41523492e-01 4.38573331e-01 5.50538123e-01 2.87275910e-01
1.03767157e+00 7.98649073e-01 -2.42261067e-01 -2.11600482e-01
-7.27288723e-01 4.98716027e-01 5.73654193e-03 6.73454881e-01
6.45941734e-01 5.75348794e-01 -2.72244960e-01 1.21751797e+00
-5.11191666e-01 3.94674093e-01 -4.15031612e-02 1.03425193e+00
9.05122086e-02 -1.38459516e+00 -3.57041448e-01 1.12646854e+00
-6.78860426e-01 -1.00340620e-01 -5.88969409e-01 4.11157072e-01
7.04357505e-01 1.26126504e+00 1.52457748e-02 -1.08272672e+00
4.50498939e-01 2.34631851e-01 4.36043024e-01 -6.37252569e-01
-1.07892156e+00 -5.43241084e-01 4.85401571e-01 -9.28619862e-01
2.39213943e-01 -7.64166713e-01 -5.14270246e-01 -5.46643496e-01
-3.14093113e-01 -3.90519738e-01 1.24975562e+00 7.72262216e-01
1.14128983e+00 6.08281970e-01 3.80381137e-01 -6.26658857e-01
-7.01366782e-01 -9.05176699e-01 -2.45321542e-01 3.13536346e-01
7.69876316e-02 -6.68774843e-01 -5.36870837e-01 -2.93929696e-01]
|
[8.790160179138184, 5.882959365844727]
|
fe4e8b72-e6fe-4a4c-b5d1-c4a2e7410a30
|
design-and-comparison-of-two-linear
|
2203.01409
| null |
https://arxiv.org/abs/2203.01409v1
|
https://arxiv.org/pdf/2203.01409v1.pdf
|
Design and comparison of two linear controllers with precompensation gain for the Quadruple inverted pendulum
|
In this work we present a workflow for designing two linear control techniques applied to the dynamic system quadruple inverted pendulum mounted on a cart (QIP) where the steady state error on cart position is eliminated through a precompensation gain. The first control law designed was based on LQR, technique that stabilizes the system states based on a minimization of the total energy of the system states and the second control technique was pole placement by 2nd order system approximation, this method allowed to define a settling time and overshoot of the desired response of the plant, for this control the system was considered SISO, being the only output of the plant the horizontal position of the cart. To eliminate the steady-state error generated by both linear controllers, an analysis of the stationary response of the system under control in the Laplace Domain was performed. The simulation environment used to validate the results presented in this paper was Simscape of Matlab, it is a simulation environment that allowed to simulate the dynamic of nonlinear mechanical systems. The results obtained indicate that both controllers are able to stabilize the plant and move the cart position to the desired setpoint. Finally, we discuss about the performances of both linear control techniques.
|
['Franklin Josue Ticona Coaquira']
|
2022-02-28
| null | null | null | null |
['total-energy']
|
['miscellaneous']
|
[ 5.03490604e-02 5.88473260e-01 1.57563627e-01 5.12389362e-01
4.73852634e-01 -8.31135452e-01 4.52761382e-01 2.20938772e-01
-2.78623074e-01 9.87371802e-01 -4.35265005e-01 -3.47892314e-01
-5.03347993e-01 -5.08720577e-01 -6.46766663e-01 -5.75901151e-01
3.17797363e-01 3.70134145e-01 1.45524502e-01 -6.81728303e-01
4.39123139e-02 5.02368271e-01 -1.23459315e+00 -4.95298445e-01
6.15837872e-01 4.74610001e-01 2.28030711e-01 6.53732955e-01
6.16242528e-01 5.61883271e-01 -3.01088363e-01 7.28133321e-01
3.76822740e-01 -4.80464250e-01 -3.45075428e-01 1.91595063e-01
-3.20924878e-01 5.20923994e-02 1.84371963e-01 1.10031748e+00
3.99081498e-01 4.24645543e-01 8.49657059e-01 -1.19290113e+00
-1.51389033e-01 3.12773049e-01 -3.26288611e-01 8.26110840e-02
3.70899767e-01 8.28989223e-02 4.01133955e-01 -2.44778290e-01
7.85339236e-01 1.24722469e+00 5.39079428e-01 6.88162297e-02
-1.61405993e+00 -1.95486173e-01 -2.50779569e-01 -3.26407179e-02
-1.39642262e+00 1.66063279e-01 2.99218297e-01 -7.80124784e-01
3.72597396e-01 4.61492270e-01 6.13419831e-01 3.76558363e-01
8.95373285e-01 7.25101009e-02 1.19583201e+00 -4.68985975e-01
1.31104872e-01 7.33534038e-01 3.73920918e-01 3.93352956e-01
4.05699998e-01 1.33177802e-01 3.79815966e-01 -7.07956627e-02
7.66950369e-01 -3.28880519e-01 -3.40562880e-01 -4.88529116e-01
-7.21135557e-01 6.58142507e-01 5.18100619e-01 8.01209033e-01
-6.64894640e-01 -2.81794071e-01 3.75676692e-01 3.52468222e-01
-1.51070207e-01 4.43940729e-01 -3.96327704e-01 1.39132977e-01
-2.09947318e-01 4.19151485e-01 1.05194998e+00 7.62097657e-01
1.05638750e-01 3.59038085e-01 1.64985985e-01 -1.83550254e-01
2.23658219e-01 6.75279379e-01 4.79040653e-01 -3.13937187e-01
1.00151617e-02 8.67656469e-01 4.05873924e-01 -1.04863477e+00
-4.97518152e-01 -4.47750866e-01 -4.48530257e-01 6.64028823e-01
4.08178985e-01 -8.03960443e-01 -3.98948997e-01 1.39752519e+00
6.46636426e-01 -5.53692102e-01 2.01842561e-01 9.43501830e-01
-1.01864599e-01 1.00194919e+00 -2.27168083e-01 -7.20934689e-01
1.26680422e+00 -7.78056458e-02 -1.10758591e+00 3.37299168e-01
1.12974435e-01 -8.92054319e-01 8.13000262e-01 3.75415415e-01
-1.02575123e+00 -7.07807481e-01 -1.37168729e+00 4.75906342e-01
-4.55006868e-01 6.90583348e-01 -7.74920285e-01 1.46783829e-01
-6.48749411e-01 9.19328928e-01 -7.85820544e-01 -3.78095955e-01
-1.01547766e+00 3.06752145e-01 -2.12980598e-01 9.85012591e-01
-1.27190983e+00 1.53726244e+00 8.70200455e-01 4.37794179e-01
4.29978035e-02 -5.48076868e-01 -3.55887294e-01 2.75615957e-02
5.15640259e-01 -7.36063361e-01 8.46287191e-01 -9.62006509e-01
-1.89443159e+00 3.10606539e-01 3.43572125e-02 -5.52206993e-01
9.83981490e-01 6.10256828e-02 7.20652789e-02 -1.61338896e-01
9.71136764e-02 -1.21055968e-01 6.63220584e-01 -8.54764521e-01
-1.69659659e-01 -4.44079526e-02 -4.15595978e-01 4.56764162e-01
2.12487265e-01 -5.27378738e-01 3.47776949e-01 -1.19506106e-01
1.47379547e-01 -1.22318494e+00 1.94529425e-02 -6.46321356e-01
-3.91522139e-01 9.51133594e-02 9.47518528e-01 -4.29356635e-01
1.07728243e+00 -2.11118054e+00 5.65479636e-01 3.90032023e-01
-6.31308377e-01 2.70661026e-01 8.09747875e-01 8.73319924e-01
-1.24658950e-01 -1.89926594e-01 2.19410323e-02 6.41341150e-01
-1.03738822e-01 -3.24144036e-01 -2.38536164e-01 5.16097963e-01
-1.60996022e-03 9.90069751e-03 -4.71497029e-01 -4.92297597e-02
4.46920484e-01 5.27908564e-01 2.28131423e-03 -2.95448378e-02
-7.81768560e-02 6.54860437e-01 -5.70537508e-01 -2.57587761e-01
4.03794706e-01 4.32692736e-01 5.67182481e-01 -3.30763847e-01
-7.98537433e-01 -2.62789756e-01 -1.58155084e+00 3.98562104e-01
-3.74864459e-01 4.09037530e-01 5.71802855e-01 -7.64642358e-01
1.19000375e+00 6.31129444e-01 5.75908244e-01 -2.50738978e-01
7.91536331e-01 4.40336108e-01 1.42057642e-01 -4.03263003e-01
-1.08325914e-01 -3.15981030e-01 5.52775085e-01 -5.47475666e-02
-3.20534378e-01 -6.91731036e-01 4.13857043e-01 -1.58209011e-01
2.55531549e-01 2.25959823e-01 7.66625702e-01 -9.29567277e-01
1.27186763e+00 3.63943964e-01 2.51603186e-01 -2.51270711e-01
1.10399768e-01 -2.62812346e-01 7.97518849e-01 8.96945223e-02
-1.12757957e+00 -1.81002617e-01 -3.48160863e-01 3.38902652e-01
1.82668269e-01 3.94333392e-01 -6.99342847e-01 4.62727427e-01
3.32938284e-01 7.44316757e-01 -2.58744180e-01 -2.90792853e-01
-6.17242932e-01 -3.87988836e-01 -1.98059782e-01 -2.14714304e-01
1.59574017e-01 -5.43218255e-01 -9.82794940e-01 4.82747585e-01
2.78750956e-01 -8.30664694e-01 1.98764116e-01 7.03059584e-02
-1.03600037e+00 -1.49567163e+00 -3.98971647e-01 -6.99169517e-01
6.25511825e-01 -4.84650999e-01 2.21091509e-01 -6.24129660e-02
-3.50227177e-01 1.94411710e-01 -8.41827393e-02 -5.42870700e-01
-7.63440967e-01 5.10745421e-02 4.36113805e-01 -1.34310767e-01
-4.07776713e-01 -9.68998894e-02 -2.11066887e-01 5.96557438e-01
-5.64996123e-01 -4.62872125e-02 2.09494844e-01 3.64282221e-01
2.85007060e-01 3.23667705e-01 5.77422857e-01 -4.33869630e-01
1.12309504e+00 -2.05274060e-01 -1.50707114e+00 -1.90354317e-01
-5.64355969e-01 -2.57857069e-02 1.14928865e+00 -3.57665688e-01
-6.89045608e-01 4.70924050e-01 8.15789253e-02 3.94033045e-02
1.51565611e-01 1.94849968e-01 -2.03404040e-03 -6.04563858e-03
6.31973565e-01 1.31443009e-01 6.94737136e-01 -4.24554944e-01
1.10102512e-01 5.77575326e-01 2.08685890e-01 -4.77916636e-02
6.67965949e-01 -1.55722454e-01 8.51020813e-01 -9.00524676e-01
1.03876203e-01 -2.77118891e-01 -5.70111334e-01 -3.96946073e-01
6.62848890e-01 -3.83623302e-01 -1.38410008e+00 3.14401358e-01
-9.92941678e-01 -8.72809440e-02 -1.13759488e-01 6.35909796e-01
-7.86499262e-01 -4.62614931e-02 -4.74759370e-01 -1.46270704e+00
-4.99676764e-01 -1.19397557e+00 2.57413030e-01 4.14163858e-01
-2.17041180e-01 -8.62367034e-01 1.72049195e-01 -3.23110849e-01
2.14581743e-01 8.24096680e-01 7.81195521e-01 -2.83910464e-02
-2.00540394e-01 -5.38311064e-01 4.77134526e-01 2.88477361e-01
-1.56722829e-01 2.91177839e-01 -2.49458700e-01 -6.99766457e-01
4.90922540e-01 9.71663296e-02 -1.27084538e-01 6.05642259e-01
-4.72015709e-01 -2.52336115e-01 -4.35911626e-01 -7.92784616e-02
2.03434515e+00 6.87639952e-01 -2.31966618e-02 5.21874607e-01
2.25046188e-01 6.44542575e-01 9.27123129e-01 8.49295929e-02
-1.76381588e-01 9.24822628e-01 2.37291649e-01 1.90939978e-02
3.65641296e-01 9.75244045e-02 3.70234877e-01 2.65396327e-01
1.21703986e-02 8.70138481e-02 -6.56834900e-01 3.01302642e-01
-1.79605520e+00 -3.53210896e-01 -8.99785340e-01 2.30864859e+00
4.14138347e-01 2.09148556e-01 4.17308688e-01 8.17855418e-01
1.13156855e+00 -6.25650585e-01 -3.17622244e-01 -7.25073099e-01
2.50143439e-01 -4.02707249e-01 9.10963356e-01 1.21678615e+00
-7.06482947e-01 1.45953968e-01 5.11899900e+00 1.31766990e-01
-1.58507097e+00 -2.65746891e-01 -1.97912276e-01 7.04482347e-02
7.24101245e-01 -1.60600133e-02 -7.27618992e-01 7.40058184e-01
9.59663212e-01 -7.85658062e-01 3.67332816e-01 9.43058670e-01
9.17719901e-01 -3.69945794e-01 -5.93764186e-01 -1.43337448e-03
-3.44575286e-01 -5.09157777e-01 -6.08910322e-01 -1.82428882e-01
7.29555845e-01 -6.39784157e-01 -7.83983096e-02 5.10402769e-02
-2.88821995e-01 -2.09275976e-01 7.43010581e-01 5.58335483e-01
5.40060289e-02 -6.88989043e-01 8.52677166e-01 9.77233112e-01
-1.06880009e+00 -4.24961776e-01 -3.26042444e-01 -4.01188403e-01
5.25224388e-01 3.84122044e-01 -1.09768796e+00 7.64548361e-01
-2.67324805e-01 1.31014392e-01 -2.35450104e-01 9.09338117e-01
-1.52145669e-01 1.49768040e-01 -6.19004130e-01 -7.18755186e-01
2.57350504e-01 -7.47399867e-01 1.08220077e+00 4.75743473e-01
-4.83360887e-03 2.56893951e-02 8.30115154e-02 1.01657093e+00
9.16290581e-01 4.89213020e-01 -6.41767919e-01 3.71698320e-01
6.10070042e-02 7.91565955e-01 -4.89992946e-01 -2.73456872e-01
6.94809616e-01 3.82322907e-01 -4.22358513e-01 1.86648384e-01
-7.43520021e-01 -8.49682510e-01 1.09731212e-01 4.59687054e-01
-2.02117354e-01 -3.16329509e-01 -2.30891228e-01 -5.53874612e-01
2.27032416e-02 -3.64076674e-01 -6.80031329e-02 -6.10085547e-01
-2.65202403e-01 3.59195679e-01 2.21327782e-01 -1.43857419e+00
-9.43691790e-01 -5.55227816e-01 -2.82854557e-01 1.35783947e+00
-6.31692290e-01 -2.24869117e-01 2.75272042e-01 2.16525525e-01
2.61724979e-01 1.69243142e-01 5.17871916e-01 2.50756353e-01
-6.25079095e-01 -5.40328026e-01 6.93983197e-01 -5.32683015e-01
3.68016183e-01 -1.20829988e+00 -3.92679334e-01 7.49507368e-01
-9.56498325e-01 6.64518595e-01 1.48196566e+00 -6.51177526e-01
-1.27233589e+00 -4.36515629e-01 6.88309729e-01 2.19507158e-01
7.53153086e-01 -5.28747356e-03 -7.89045155e-01 6.52169049e-01
1.60794601e-01 -4.00842041e-01 -4.71556991e-01 -7.91233063e-01
8.95155907e-01 -2.77640522e-01 -1.20515943e+00 3.24953586e-01
-4.38519806e-01 3.78300548e-02 -5.76541007e-01 3.30522090e-01
3.55085641e-01 -7.68672824e-01 -8.07584405e-01 2.33192056e-01
3.36552799e-01 -6.50965512e-01 4.68481034e-01 -1.36619478e-01
-1.58429727e-01 -9.35219586e-01 6.65936410e-01 -1.45630729e+00
-1.16343506e-01 -8.02652359e-01 9.55964848e-02 8.46181214e-01
3.08144897e-01 -6.73828125e-01 1.86805308e-01 2.99603373e-01
5.57905078e-01 -5.93407989e-01 -7.59059906e-01 -5.06434977e-01
-1.82809770e-01 6.92193866e-01 -2.05422238e-01 4.93735224e-01
4.36287880e-01 6.33872867e-01 -1.08111285e-01 6.74583137e-01
2.48329669e-01 -1.99877307e-01 6.87443674e-01 -1.01742363e+00
-1.34037971e-01 -1.72756925e-01 -5.94592810e-01 -1.49673477e-01
-1.44773334e-01 -1.60924822e-01 -1.07641920e-01 -1.63289583e+00
-6.99994683e-01 1.76289737e-01 9.76005718e-02 -3.67240123e-02
-7.27307871e-02 -5.68870381e-02 2.36382097e-01 2.17198908e-01
6.79176569e-01 2.41284985e-02 1.10850954e+00 2.95315385e-01
-9.62172329e-01 7.57608593e-01 2.41982430e-01 4.94911343e-01
6.50831103e-01 -1.66373685e-01 -9.55281973e-01 2.28485137e-01
8.58059749e-02 3.40558827e-01 -6.84481561e-02 -1.41282499e+00
-1.04948254e-02 -4.84793112e-02 3.47167253e-02 -6.36972964e-01
4.29412536e-02 -1.33811402e+00 1.00356412e+00 1.30983329e+00
-3.18103842e-02 4.68566567e-02 1.85008198e-01 2.33934835e-01
-1.32449076e-01 -4.91834432e-01 1.23716104e+00 2.32195675e-01
-3.11344247e-02 -7.38331914e-01 -6.73468411e-01 -7.31296301e-01
1.53286314e+00 -1.50791138e-01 1.49757236e-01 -3.72594535e-01
-9.16022718e-01 4.56858426e-01 2.31913358e-01 1.37483589e-02
-5.77248394e-01 -6.51333869e-01 -3.13701302e-01 1.20328330e-01
-7.95596898e-01 -4.81943250e-01 3.23128045e-01 1.03931773e+00
-1.04564488e+00 9.33548152e-01 -7.99623370e-01 -4.08176750e-01
-1.58726072e+00 8.04608762e-01 8.35244358e-01 3.31518543e-03
-3.73433381e-01 -9.92629305e-02 -5.65878510e-01 -2.45580822e-01
-1.00949913e-01 -5.84351957e-01 -6.26105011e-01 2.63853490e-01
-7.08051696e-02 6.05756760e-01 1.47425875e-01 -5.72443366e-01
-1.51406288e-01 7.58899271e-01 6.94748044e-01 -4.64965373e-01
1.10416389e+00 -6.96890354e-02 -1.37634084e-01 6.99603200e-01
7.73864567e-01 -8.67570639e-02 -1.00649381e+00 6.43444836e-01
-1.09983444e-01 7.92815723e-03 -1.07606605e-01 -7.75673270e-01
-3.10746014e-01 2.88051933e-01 7.20961273e-01 9.54187751e-01
7.71900475e-01 -1.22863066e+00 -5.53208105e-02 3.97900641e-01
2.41494551e-01 -1.04096818e+00 -7.18766272e-01 5.38995564e-01
9.51726735e-01 -4.49978709e-01 3.58132035e-01 -8.45637321e-01
-7.65702844e-01 1.41656518e+00 4.80743080e-01 -7.33380318e-01
4.53859270e-01 2.19767511e-01 -1.24798134e-01 4.47840869e-01
-3.58761787e-01 -1.55591965e-01 1.67757779e-01 1.49838310e-02
6.52403295e-01 2.44719312e-02 -1.59089303e+00 2.37166449e-01
1.96864624e-02 6.35960996e-01 9.98707056e-01 9.89577472e-01
-6.93841279e-01 -8.18675876e-01 -1.07286680e+00 -3.26710820e-01
-4.51661617e-01 6.65292978e-01 -2.72854328e-01 1.41230929e+00
2.50781178e-01 6.29961967e-01 -5.71141578e-02 2.62751251e-01
1.32399070e+00 1.96801886e-01 1.80591241e-01 -3.43374074e-01
-7.26448059e-01 1.23330548e-01 -1.08889081e-01 -2.16065161e-02
1.46137282e-01 -3.16981256e-01 -1.34871113e+00 -1.05816431e-01
-6.11965358e-01 4.92489547e-01 1.15702057e+00 5.52055359e-01
1.27511069e-01 6.56930208e-01 5.48609614e-01 -6.18929267e-01
-9.93535221e-01 -1.13826120e+00 -7.15641379e-01 -6.70539215e-02
4.47698176e-01 -8.54135275e-01 -5.14996827e-01 9.51971626e-04]
|
[5.441054821014404, 2.559298038482666]
|
d615b6fb-121b-4284-9afd-4ad62264f9aa
|
distance-aware-occlusion-detection-with
|
2208.11122
| null |
https://arxiv.org/abs/2208.11122v1
|
https://arxiv.org/pdf/2208.11122v1.pdf
|
Distance-Aware Occlusion Detection with Focused Attention
|
For humans, understanding the relationships between objects using visual signals is intuitive. For artificial intelligence, however, this task remains challenging. Researchers have made significant progress studying semantic relationship detection, such as human-object interaction detection and visual relationship detection. We take the study of visual relationships a step further from semantic to geometric. In specific, we predict relative occlusion and relative distance relationships. However, detecting these relationships from a single image is challenging. Enforcing focused attention to task-specific regions plays a critical role in successfully detecting these relationships. In this work, (1) we propose a novel three-decoder architecture as the infrastructure for focused attention; 2) we use the generalized intersection box prediction task to effectively guide our model to focus on occlusion-specific regions; 3) our model achieves a new state-of-the-art performance on distance-aware relationship detection. Specifically, our model increases the distance F1-score from 33.8% to 38.6% and boosts the occlusion F1-score from 34.4% to 41.2%. Our code is publicly available.
|
['Guyue Zhou', 'Hao Zhao', 'Xiaoxue Chen', 'Yucheng Tu', 'Yang Li']
|
2022-08-23
| null | null | null | null |
['visual-relationship-detection']
|
['computer-vision']
|
[ 2.67743230e-01 -4.38957736e-02 -2.41098046e-01 -7.55650520e-01
-6.28281653e-01 -3.99128467e-01 4.98741060e-01 4.25728083e-01
-3.49284142e-01 2.38249823e-01 2.87210166e-01 -1.19755268e-01
5.11159152e-02 -5.87784290e-01 -7.89569020e-01 -9.75028053e-02
-1.51387021e-01 4.64400828e-01 6.29478157e-01 -2.58050382e-01
3.63627195e-01 5.90123594e-01 -1.44190323e+00 3.93334031e-01
8.06782722e-01 8.53857934e-01 3.67132366e-01 6.84951186e-01
1.93838790e-01 6.68710172e-01 -4.87222135e-01 -1.19554356e-01
-1.50672887e-02 -2.06023589e-01 -7.54656494e-01 -6.60078749e-02
8.59525919e-01 -4.85199600e-01 -3.83763582e-01 9.52534258e-01
2.88123369e-01 2.57557809e-01 6.17860973e-01 -1.35712802e+00
-8.08881938e-01 4.06729698e-01 -1.22861660e+00 4.05168414e-01
2.74985760e-01 1.65057614e-01 1.27189231e+00 -1.07664716e+00
4.65691566e-01 1.56283224e+00 1.12470515e-01 1.95501670e-01
-1.21761382e+00 -6.92569673e-01 5.83499134e-01 4.46949422e-01
-1.51016557e+00 -3.43334079e-01 6.53084457e-01 -5.00995636e-01
1.31650245e+00 2.44342297e-01 5.48472703e-01 5.25756478e-01
-2.20942393e-01 9.26952660e-01 5.76581717e-01 -4.55426306e-01
-2.20521972e-01 -1.47537231e-01 3.08140874e-01 5.90978026e-01
3.04023087e-01 -1.67463720e-01 -6.77744329e-01 2.49774978e-01
9.05497253e-01 -2.47815132e-01 -1.52696088e-01 -3.80388498e-01
-1.16851115e+00 6.89205110e-01 1.00736749e+00 2.25505918e-01
-1.17970534e-01 4.47566241e-01 1.55300364e-01 -1.87388912e-01
4.87952739e-01 4.47915047e-01 -2.06184283e-01 1.67264678e-02
-5.83575785e-01 2.91611642e-01 3.61903667e-01 1.08559322e+00
5.87864280e-01 -3.91678751e-01 -3.46507430e-01 1.06259036e+00
4.31559831e-01 3.72524291e-01 -2.28964493e-01 -7.63361156e-01
7.16785550e-01 5.99752724e-01 1.09863915e-01 -1.40530491e+00
-6.48971558e-01 -5.31353474e-01 -4.31640476e-01 2.02387497e-02
4.25487161e-01 3.38423193e-01 -8.01128685e-01 1.78059781e+00
3.30790311e-01 -8.77826214e-02 -2.96817690e-01 1.38096166e+00
9.19802547e-01 6.27828598e-01 3.28363419e-01 2.47429445e-01
1.65966785e+00 -1.16539633e+00 -5.98100066e-01 -7.85983264e-01
3.86085093e-01 -8.58060122e-01 1.03751838e+00 2.89551876e-02
-1.01274145e+00 -7.62905359e-01 -1.07791483e+00 -6.54902816e-01
-1.23474687e-01 2.60402113e-01 8.68950188e-01 5.81263304e-02
-9.01448607e-01 1.75498202e-01 -9.04086888e-01 -5.57614267e-01
7.41974294e-01 5.56082964e-01 -2.24487826e-01 -2.26127654e-01
-8.75545025e-01 9.01115596e-01 2.61517256e-01 2.85454150e-02
-5.13685048e-01 -4.72234458e-01 -1.05729830e+00 2.55610913e-01
6.53731704e-01 -7.04447031e-01 1.11567974e+00 -6.09408140e-01
-6.76673830e-01 1.11749554e+00 -5.63517690e-01 -2.71820635e-01
3.29522192e-01 -6.53783321e-01 -1.34749204e-01 2.15760931e-01
3.12356830e-01 1.20036650e+00 3.84909004e-01 -1.51967084e+00
-6.87950969e-01 -4.17808294e-01 1.52699366e-01 4.12397116e-01
-2.26822197e-01 4.53793138e-01 -9.90079939e-01 -7.65960813e-01
4.09975827e-01 -9.16828275e-01 -9.94009301e-02 6.59927428e-01
-4.70893621e-01 -3.74485970e-01 7.90206134e-01 -6.41147971e-01
9.07331705e-01 -2.08991599e+00 7.02873766e-02 1.01410359e-01
5.80265284e-01 1.69089064e-01 -1.49469256e-01 -6.50458783e-02
-7.02567212e-03 5.80868125e-03 -1.54198930e-01 -3.61440212e-01
-1.23629183e-01 -7.04396740e-02 -9.97956991e-02 3.01605225e-01
5.99390566e-01 1.20145905e+00 -9.41413999e-01 -4.83045608e-01
3.30497116e-01 6.17539465e-01 -6.33132339e-01 2.20478833e-01
-2.33893558e-01 4.41265345e-01 -3.05208445e-01 6.51360154e-01
6.77491367e-01 -4.63512391e-01 7.65032247e-02 -5.42289197e-01
3.17871459e-02 3.62185270e-01 -8.91263247e-01 1.74114835e+00
-3.32140446e-01 1.22220516e+00 -2.43871704e-01 -9.04119730e-01
1.12644994e+00 -2.33831465e-01 2.99777418e-01 -9.41504657e-01
1.06059395e-01 -1.50985092e-01 3.53942245e-01 -5.60989559e-01
5.82283616e-01 2.84837902e-01 1.05932765e-01 1.37231126e-01
-3.67591411e-01 7.33632967e-02 1.52779862e-01 2.84425229e-01
8.03915977e-01 2.61978298e-01 4.58179742e-01 -8.61505345e-02
3.96134943e-01 -8.31533037e-03 3.31916064e-01 5.31188607e-01
-2.90225714e-01 7.95482635e-01 4.87345904e-01 -3.59019220e-01
-7.90834546e-01 -1.17391729e+00 3.15307304e-02 1.15374672e+00
6.87315643e-01 -4.51349795e-01 -2.69369125e-01 -4.66572315e-01
1.88845068e-01 7.05780327e-01 -5.21425843e-01 -2.67063290e-01
-9.20020401e-01 -4.89024967e-01 5.14524654e-02 1.00259709e+00
6.42692864e-01 -7.96515405e-01 -6.05241716e-01 -3.61049250e-02
-4.10146654e-01 -1.65690637e+00 -6.49431348e-01 -7.37714320e-02
-6.23701572e-01 -1.02559376e+00 -5.18329263e-01 -9.62507367e-01
7.87506640e-01 7.03973532e-01 1.16480553e+00 6.16768412e-02
-5.54019213e-01 -3.12699266e-02 -1.02855735e-01 -4.59365338e-01
1.67572752e-01 3.42566939e-03 -2.51401424e-01 -1.01297595e-01
5.17989874e-01 -3.25610548e-01 -7.53316045e-01 6.97821856e-01
-2.87125945e-01 2.72622645e-01 5.55195749e-01 3.51075888e-01
6.75364435e-01 -2.25018352e-01 3.35218728e-01 -5.90372622e-01
2.25111574e-01 -2.60278940e-01 -5.77287555e-01 1.82633057e-01
-2.13335112e-01 2.80710054e-03 2.90101409e-01 -3.39739084e-01
-9.45570409e-01 7.61311874e-02 1.09681688e-01 -3.65876257e-01
-1.13492370e-01 1.30462868e-03 -2.42771700e-01 7.17714243e-03
6.75324738e-01 -1.71253458e-01 -3.56290460e-01 -3.64206761e-01
5.46871126e-01 5.32645226e-01 7.03554571e-01 -4.89700377e-01
6.63532853e-01 6.34081721e-01 8.03726315e-02 -5.32227993e-01
-1.10463107e+00 -6.97689056e-01 -8.16721022e-01 -2.98065186e-01
1.01316357e+00 -9.75422859e-01 -8.62141788e-01 7.24816993e-02
-1.39779413e+00 -2.92135179e-01 2.29068428e-01 5.53309262e-01
-1.90798074e-01 2.76873708e-01 -3.70136857e-01 -6.41142964e-01
-1.80074379e-01 -1.12790883e+00 1.35893345e+00 3.01372766e-01
-4.63683814e-01 -6.99276447e-01 -3.52837205e-01 6.17536724e-01
2.76186198e-01 2.29081795e-01 8.20836127e-01 -3.00855458e-01
-8.43977034e-01 1.20427527e-01 -1.02067471e+00 -9.16178226e-02
1.92375109e-01 -1.15978085e-01 -1.00374699e+00 -4.09787036e-02
-4.69215631e-01 -1.43839911e-01 1.02002645e+00 4.90301132e-01
1.35718620e+00 2.18330637e-01 -8.39403450e-01 4.84025002e-01
1.03551459e+00 3.33580077e-01 3.70038688e-01 2.91622877e-01
1.06440592e+00 8.74739707e-01 8.72759700e-01 2.53075570e-01
7.38706052e-01 1.31226027e+00 5.64852953e-01 -3.42412412e-01
-5.16661465e-01 -2.54665107e-01 -2.19747797e-01 1.78857461e-01
4.09221724e-02 -2.39293307e-01 -1.04360676e+00 6.21913970e-01
-2.10898423e+00 -7.80441105e-01 -5.65091848e-01 2.11750507e+00
7.50926733e-01 5.26838779e-01 1.65609345e-01 -8.84113908e-02
9.33381081e-01 3.78858037e-02 -5.72952628e-01 -1.66505441e-01
7.31744543e-02 -1.60959587e-01 2.74790764e-01 6.15794718e-01
-1.36702788e+00 1.21301711e+00 5.64731789e+00 4.66709942e-01
-9.02971208e-01 -2.23276734e-01 6.90223098e-01 -2.17411190e-01
-1.01162866e-01 6.02299580e-03 -9.94058132e-01 1.53463662e-01
3.02433759e-01 7.51348287e-02 1.56424612e-01 7.65534759e-01
2.74262816e-01 -4.99396890e-01 -1.36033726e+00 1.31992447e+00
3.71651709e-01 -9.66915548e-01 4.64171246e-02 -6.27982169e-02
4.81382489e-01 -3.09108138e-01 8.66048634e-02 4.23264392e-02
-4.18906510e-02 -1.21993983e+00 7.67981052e-01 3.72548044e-01
7.60152280e-01 -6.68160856e-01 2.88139731e-01 6.64096773e-02
-1.55895483e+00 1.83684886e-01 -2.58379728e-01 -1.44784838e-01
3.46285939e-01 6.31133735e-01 -8.65962625e-01 2.11544782e-01
7.56089866e-01 9.65593874e-01 -7.70470500e-01 1.15879583e+00
-4.91449326e-01 2.76718438e-01 -4.41807300e-01 3.48685645e-02
-2.15787031e-02 1.19047068e-01 4.00421828e-01 1.15616131e+00
-1.88763067e-01 1.86125323e-01 1.86298251e-01 9.80270028e-01
-9.72012505e-02 -8.05234760e-02 -4.65282947e-01 3.28028589e-01
4.46736813e-01 1.09946525e+00 -9.80877936e-01 -1.92587972e-01
-4.89177942e-01 1.00440180e+00 6.44231141e-01 1.81433290e-01
-1.16367686e+00 -4.48905945e-01 7.75953054e-01 1.59309521e-01
3.70932192e-01 -3.97889972e-01 -5.63723803e-01 -8.22789490e-01
3.53769213e-01 -3.56063575e-01 3.31147790e-01 -1.10319805e+00
-1.05505538e+00 4.03112352e-01 2.32378155e-01 -1.02858949e+00
7.22868964e-02 -5.33396959e-01 -2.63002425e-01 8.32679570e-01
-1.62478399e+00 -1.13400078e+00 -7.72102535e-01 1.76386386e-01
7.36266077e-01 2.70396113e-01 4.02575284e-01 4.70647395e-01
-5.05766690e-01 6.94948912e-01 -5.62332749e-01 3.55220616e-01
7.36550570e-01 -9.99315143e-01 8.25767159e-01 8.76399398e-01
3.79176855e-01 5.83341897e-01 5.34202099e-01 -6.32361889e-01
-1.03298593e+00 -1.06453395e+00 1.05830753e+00 -4.93801266e-01
4.37734544e-01 -5.57294130e-01 -1.02090049e+00 7.41168678e-01
-1.16766609e-01 2.15250105e-01 3.75820100e-01 2.75458992e-01
-7.39554644e-01 -1.11738607e-01 -8.94356608e-01 8.76297891e-01
1.46939528e+00 -4.97137219e-01 -3.70700598e-01 3.60306501e-01
7.36904800e-01 -6.15921915e-01 -5.58297873e-01 4.43380862e-01
6.35463238e-01 -7.64041722e-01 1.35222948e+00 -3.30929965e-01
5.16982853e-01 -5.35482883e-01 -2.04710752e-01 -7.95408666e-01
-5.29065788e-01 -2.99549907e-01 -2.11361468e-01 1.00847507e+00
3.40125680e-01 -2.92297781e-01 5.99320054e-01 6.45978510e-01
2.67409999e-02 -8.61525893e-01 -5.05558074e-01 -5.86767137e-01
-3.61019313e-01 -5.59716344e-01 1.23163126e-01 6.79484844e-01
-1.65205583e-01 7.80401886e-01 -2.17420697e-01 3.87826264e-01
4.95556176e-01 3.24440867e-01 8.24406624e-01 -1.07697499e+00
-3.44401926e-01 -5.21168232e-01 -5.15451550e-01 -1.70961392e+00
-3.06804497e-02 -8.12886953e-01 2.49021165e-02 -1.86204386e+00
4.94551450e-01 -5.01252711e-01 -1.82242021e-02 5.77189922e-01
-5.66017687e-01 4.97723192e-01 3.60230237e-01 1.79272547e-01
-7.80424356e-01 3.66802901e-01 1.32728279e+00 -2.16915578e-01
-1.04092628e-01 -1.62129983e-01 -9.34800982e-01 6.14289224e-01
7.92705953e-01 -3.53578031e-01 -4.59564894e-01 -7.03390419e-01
4.31871116e-02 -2.34437168e-01 6.41980469e-01 -9.12794292e-01
1.96421698e-01 -1.17852196e-01 4.56254005e-01 -7.86442757e-01
6.13521814e-01 -6.60627782e-01 -4.16661441e-01 3.46693367e-01
-4.73615766e-01 -9.38852951e-02 4.15118247e-01 5.24638474e-01
-7.03222081e-02 2.14939173e-02 6.84014201e-01 1.91682532e-01
-1.07144499e+00 9.02067572e-02 1.63592100e-01 2.89250433e-01
1.11461103e+00 -1.36080608e-01 -4.67623949e-01 -5.19271433e-01
-5.58074296e-01 5.49363971e-01 1.84914112e-01 6.17818713e-01
7.43646801e-01 -1.12979746e+00 -7.63242304e-01 -1.20276986e-02
4.48811859e-01 2.64610350e-01 2.28704978e-02 6.20918930e-01
-4.64230567e-01 4.89765257e-01 -8.99507999e-02 -9.07289386e-01
-1.60738850e+00 4.98947561e-01 3.02605391e-01 2.53125310e-01
-6.88551962e-01 1.25259590e+00 5.73415816e-01 2.32352093e-01
3.80405664e-01 -4.32335645e-01 -3.94534647e-01 -1.63260370e-01
5.62398970e-01 1.56539947e-01 -1.87187597e-01 -7.70241439e-01
-7.05526769e-01 1.13798332e+00 -2.77382940e-01 9.37570632e-02
1.10430706e+00 -2.74074394e-02 -5.45427902e-03 3.74134898e-01
1.38076460e+00 -1.27726272e-01 -1.36514735e+00 -3.86182338e-01
-7.18299374e-02 -8.25218439e-01 -3.36051881e-02 -8.36035907e-01
-1.08489048e+00 9.54561651e-01 3.18272263e-01 -6.31043920e-03
1.01756454e+00 5.41826129e-01 6.19911671e-01 1.12259455e-01
1.13186903e-01 -6.58567131e-01 4.42446500e-01 3.43299448e-01
1.00415599e+00 -1.66741490e+00 2.97089994e-01 -1.19421411e+00
-6.31230295e-01 9.34701622e-01 1.08989775e+00 -2.72901766e-02
4.17054892e-01 1.87054902e-01 -2.48004898e-01 -4.65463698e-01
-6.34041011e-01 -5.23761094e-01 8.05634797e-01 5.77009022e-01
7.30906904e-01 5.23493066e-02 -1.30682558e-01 2.39513472e-01
-1.36391163e-01 -3.86215717e-01 2.86217686e-02 6.92676783e-01
-5.12160242e-01 -7.62892723e-01 -3.31337631e-01 3.58711839e-01
-1.82185009e-01 -1.83455899e-01 -3.50199401e-01 6.86212599e-01
-7.80512989e-02 1.14087057e+00 5.19061625e-01 -1.71019211e-01
5.34882784e-01 -3.69063169e-01 5.75538814e-01 -7.12081969e-01
-2.47803152e-01 8.47181603e-02 1.95582584e-01 -5.21121621e-01
-4.07761306e-01 -5.11498511e-01 -1.68401730e+00 -6.41558990e-02
-2.63170004e-01 -4.19645846e-01 5.82329333e-01 8.35396051e-01
4.21326429e-01 8.10108364e-01 3.64186168e-01 -7.28132129e-01
6.73600957e-02 -7.15193212e-01 -3.38815302e-02 4.69900668e-01
2.05868751e-01 -9.99715388e-01 2.25533731e-02 -3.03162076e-02]
|
[10.207860946655273, 1.5828046798706055]
|
f3aad835-03a3-425e-81ea-d13ee7cfdd69
|
delidata-a-dataset-for-deliberation-in-multi
|
2108.05271
| null |
https://arxiv.org/abs/2108.05271v3
|
https://arxiv.org/pdf/2108.05271v3.pdf
|
DeliData: A dataset for deliberation in multi-party problem solving
|
Group deliberation enables people to collaborate and solve problems, however, it is understudied due to a lack of resources. To this end, we introduce the first publicly available dataset containing collaborative conversations on solving a well-established cognitive task, consisting of 500 group dialogues and 14k utterances. In 64% of these conversations, the group members are able to find a better solution than they had identified individually, and in 43.8% of the groups who had a correct answer as their final solution, none of the participants had solved the task correctly by themselves. Furthermore, we propose a novel annotation schema that captures deliberation cues and release all 14k utterances annotated with it. Finally, we use the proposed dataset to develop and evaluate two methods for generating deliberation utterances. The data collection platform, dataset and annotated corpus are publicly available at https://delibot.xyz.
|
['Andreas Vlachos', 'Tom Stafford', 'Georgi Karadzhov']
|
2021-08-11
| null | null | null | null |
['problem-solving-deliberation']
|
['natural-language-processing']
|
[ 1.43334493e-02 8.76261413e-01 2.97893137e-01 -5.76828480e-01
-8.41475248e-01 -6.75015032e-01 7.88697958e-01 2.54109830e-01
-3.63208473e-01 1.12959898e+00 1.08871353e+00 7.98436180e-02
-1.03426754e-01 -5.12493610e-01 -5.34300469e-02 -4.65216696e-01
4.56284851e-01 6.98695898e-01 -3.35215002e-01 -2.55818427e-01
5.01512587e-01 -6.67711616e-01 -1.31694603e+00 7.76331663e-01
1.27987921e+00 6.45352364e-01 2.06187487e-01 7.22298324e-01
-3.36907983e-01 1.13472641e+00 -9.96414185e-01 -5.53955436e-01
3.53251304e-03 -6.78155541e-01 -1.65127766e+00 1.33426830e-01
1.99857071e-01 -2.75961310e-01 5.98365581e-03 9.25931990e-01
7.16665149e-01 3.50501150e-01 4.02037948e-01 -1.21003544e+00
-5.12909710e-01 1.35657179e+00 3.63564521e-01 -2.93075383e-01
1.14776742e+00 3.49199921e-01 9.86950874e-01 -6.56334341e-01
7.48259962e-01 1.37558389e+00 6.55463636e-01 8.90634298e-01
-9.83432949e-01 -4.53200370e-01 1.19219974e-01 2.99397945e-01
-1.07768548e+00 -7.65917361e-01 6.91763282e-01 -5.70618331e-01
8.16532135e-01 5.91442466e-01 7.57254004e-01 1.42837119e+00
-3.58547986e-01 7.33988702e-01 1.44707859e+00 -3.84734601e-01
2.43171841e-01 2.43006557e-01 5.60206711e-01 5.92799902e-01
-1.48792982e-01 -5.51961899e-01 -8.20541382e-01 -3.90115917e-01
2.22257599e-01 -3.43456537e-01 -3.90663177e-01 4.26974475e-01
-1.43286455e+00 7.02989757e-01 2.18275025e-01 4.96023208e-01
-4.54354256e-01 -2.15206534e-01 1.96061149e-01 5.33913910e-01
4.98302191e-01 8.57829452e-01 -3.13469261e-01 -6.85270607e-01
-8.30084383e-02 8.54235053e-01 1.52748382e+00 9.05750334e-01
8.73510838e-01 -6.00878119e-01 -6.52916789e-01 8.24056983e-01
2.80409157e-01 2.33092606e-01 4.61156726e-01 -1.48255682e+00
8.63488615e-01 1.12599874e+00 6.06295526e-01 -1.02624297e+00
-7.69579589e-01 5.54247126e-02 -7.53444433e-01 -4.46606517e-01
9.12691474e-01 -7.33347774e-01 4.56745364e-02 1.63215089e+00
4.23102140e-01 -2.44227096e-01 5.41094184e-01 1.06941378e+00
1.39011717e+00 3.50343794e-01 -2.20110063e-02 -3.90078813e-01
1.55105484e+00 -1.17110169e+00 -1.17344022e+00 -3.93628292e-02
9.39731658e-01 -8.43864441e-01 1.00894856e+00 2.83081770e-01
-1.06154108e+00 -4.56545562e-01 -3.42587024e-01 -6.25477284e-02
3.89857739e-02 -1.36520669e-01 5.84455013e-01 5.76958776e-01
-9.76486742e-01 2.58978784e-01 -5.13819698e-03 -4.56569582e-01
-3.50423381e-02 -1.84953645e-01 -1.18032627e-01 -3.84665914e-02
-1.39258981e+00 7.19631255e-01 3.71936083e-01 1.23977214e-01
-6.17540240e-01 -6.15558863e-01 -6.40269518e-01 -7.50140548e-02
6.87772393e-01 -7.61970758e-01 1.69673359e+00 -6.14462733e-01
-1.76005924e+00 7.76395082e-01 -2.44146049e-01 -2.96652526e-01
8.55825484e-01 -2.16175064e-01 1.82566382e-02 -1.98935315e-01
3.72652650e-01 4.06543046e-01 2.26022184e-01 -1.07663250e+00
-6.23779118e-01 -1.59009293e-01 5.34801424e-01 6.96649015e-01
-8.04867595e-02 -4.49743159e-02 8.19093138e-02 -2.94345051e-01
7.02339336e-02 -9.29253817e-01 -3.11834812e-01 -6.80692375e-01
-7.96306610e-01 -8.73869300e-01 8.61044526e-02 -7.63869464e-01
1.19382310e+00 -1.80920076e+00 3.35402310e-01 -1.11112267e-01
7.51947701e-01 2.42868468e-01 -5.90752959e-02 1.01197445e+00
4.80489016e-01 5.14452875e-01 -2.22018901e-02 -7.45684624e-01
4.87603039e-01 4.16921377e-02 -9.34747905e-02 3.52755450e-02
-4.14968193e-01 9.04706836e-01 -1.00593150e+00 -3.61119598e-01
3.93633172e-02 -6.63542822e-02 -5.43227792e-01 6.92671061e-01
-5.30922115e-01 9.10868347e-01 -6.79695249e-01 1.08278222e-01
5.23527622e-01 -2.91923851e-01 4.97858971e-01 2.45115876e-01
-3.61775070e-01 8.29480410e-01 -1.13583946e+00 1.72273457e+00
-1.97418705e-01 4.51756537e-01 4.69960012e-02 -5.51765144e-01
8.20866048e-01 6.17691636e-01 6.40229210e-02 -6.89849496e-01
2.00260162e-01 2.03019455e-02 1.62697405e-01 -8.61744165e-01
6.83148861e-01 4.72692579e-01 -6.00565195e-01 1.28109574e+00
-2.19470724e-01 -1.25215545e-01 2.64604896e-01 3.92512858e-01
1.22977138e+00 -7.42033720e-01 4.44614530e-01 -2.93010920e-02
6.42436087e-01 5.09758592e-02 5.75994849e-01 1.44263339e+00
-3.28739285e-01 3.06351900e-01 7.12619603e-01 -5.92770696e-01
-5.87865531e-01 -4.73528326e-01 4.13376033e-01 1.10498643e+00
3.91481705e-02 -9.45912004e-01 -1.09996963e+00 -6.81866527e-01
-2.42076814e-01 6.10233247e-01 -5.92564344e-01 4.19552863e-01
-4.11372155e-01 -2.04443395e-01 7.18882442e-01 -2.18916237e-01
1.27858520e+00 -1.35818112e+00 -3.63337457e-01 2.91026413e-01
-1.22497082e+00 -1.10550976e+00 -3.81070882e-01 -5.81855237e-01
-1.72328115e-01 -1.51514220e+00 -2.97676593e-01 -5.48212290e-01
5.15703440e-01 3.53002250e-01 1.07851160e+00 6.98548913e-01
2.06917822e-01 5.56702137e-01 -8.22343290e-01 -3.43343437e-01
-6.72167063e-01 2.28735104e-01 2.61737723e-02 2.67976858e-02
3.49656224e-01 -3.09435070e-01 -3.79526854e-01 4.99385238e-01
3.19965295e-02 5.94917476e-01 -7.29396716e-02 3.88963729e-01
-3.90900344e-01 2.68667229e-02 6.04578257e-01 -9.62447584e-01
1.57894528e+00 -4.19731051e-01 9.41115469e-02 1.56520635e-01
-1.98379651e-01 -2.45537907e-01 3.70008886e-01 -1.61518101e-02
-1.27681732e+00 -3.56201589e-01 -1.21300377e-01 4.50519860e-01
-4.21280324e-01 5.38545370e-01 2.08104670e-01 3.06611240e-01
7.29194999e-01 2.52943486e-02 3.37015428e-02 -5.67425430e-01
3.05046797e-01 8.75535309e-01 3.07408661e-01 -8.68101478e-01
4.20813680e-01 -1.33465767e-01 -8.75017822e-01 -6.78619802e-01
-1.16070473e+00 -2.75510907e-01 -4.94835913e-01 -7.71603882e-01
1.00193787e+00 -8.90352547e-01 -1.49941385e+00 6.82200849e-01
-1.66718686e+00 -6.97683394e-01 1.61481410e-01 2.73953408e-01
-3.94694686e-01 3.42918366e-01 -4.60672349e-01 -1.14028382e+00
-4.44850445e-01 -9.95632112e-01 4.24245000e-01 3.79866630e-01
-1.18639302e+00 -1.01916087e+00 -6.04863837e-02 1.13727498e+00
5.27667761e-01 1.38779551e-01 5.22659361e-01 -1.01499522e+00
-3.98112774e-01 1.91391230e-01 -4.73402552e-02 -1.88308328e-01
1.18055627e-01 -4.35428798e-01 -7.47970343e-01 2.38750726e-02
1.22950964e-01 -7.35867321e-01 3.87881964e-01 -4.76746298e-02
8.01034212e-01 -9.22642946e-01 -1.19453453e-01 -1.28783509e-01
3.73340011e-01 -7.53674433e-02 3.73371303e-01 9.92897153e-02
4.73429263e-01 1.15757155e+00 5.53014934e-01 8.75436246e-01
1.19271994e+00 5.24369121e-01 -3.86811644e-02 4.76168394e-01
-1.57603826e-02 -2.92809665e-01 2.07933486e-01 1.11829805e+00
-2.58700043e-01 -4.11083907e-01 -1.22532642e+00 3.58356565e-01
-2.34116554e+00 -1.05117786e+00 -4.75996435e-01 1.56044757e+00
1.16513801e+00 -2.77670741e-01 1.24738440e-01 1.40431225e-01
7.89806426e-01 2.46620983e-01 -1.27748221e-01 -2.00797319e-01
-1.04178771e-01 -1.65678769e-01 -3.60298246e-01 1.04690194e+00
-7.73364007e-01 9.44777489e-01 6.37843561e+00 2.95261651e-01
-3.78093034e-01 2.61636168e-01 4.58879352e-01 1.43523008e-01
-4.05606329e-01 -5.43806814e-02 -8.28053713e-01 6.57231748e-01
6.73779905e-01 -4.55818802e-01 6.86958313e-01 4.64131832e-01
4.41619784e-01 -3.67866188e-01 -1.26062977e+00 1.02261746e+00
7.08111748e-02 -1.46039069e+00 -3.88640285e-01 -1.20009467e-01
7.61463940e-01 -4.66264933e-01 -3.66084874e-01 5.85503876e-01
8.02168429e-01 -1.12349916e+00 7.54109204e-01 6.46464348e-01
1.20657541e-01 -2.99275368e-01 8.26128304e-01 1.11365330e+00
-7.36703515e-01 -1.65451318e-01 8.64611706e-04 -9.75552678e-01
3.77156675e-01 3.99674773e-01 -1.33302891e+00 3.74991864e-01
5.17198920e-01 5.14829695e-01 -4.93446350e-01 5.77252686e-01
-6.97426796e-01 5.60168684e-01 -1.24480702e-01 -5.58317125e-01
1.05575614e-01 -5.81029840e-02 6.30236506e-01 1.03949070e+00
9.18307602e-02 7.80849814e-01 6.82712734e-01 9.31394339e-01
-2.52088815e-01 1.17083946e-02 -2.82179683e-01 -1.17682323e-01
1.08197391e+00 1.14340484e+00 -2.70968884e-01 -4.67447281e-01
1.15763433e-01 6.25473559e-01 7.59713054e-01 3.72153729e-01
-4.93785441e-01 1.36628253e-02 7.28824973e-01 -2.47438535e-01
-5.18496931e-01 -1.16477735e-01 -4.90906656e-01 -1.27155948e+00
1.07865632e-01 -1.32281780e+00 3.17858040e-01 -7.23349452e-01
-1.38345051e+00 4.13207918e-01 -1.74704820e-01 -5.89985430e-01
-4.96149689e-01 4.82124276e-02 -7.66703308e-01 9.47207808e-01
-9.91310656e-01 -6.95437133e-01 -1.06487203e+00 3.91601831e-01
6.09714925e-01 -2.91167587e-01 1.05860782e+00 2.34331656e-02
-6.46635532e-01 3.58204901e-01 -6.85880601e-01 2.89581984e-01
9.02084172e-01 -1.04644716e+00 2.28533790e-01 2.05839559e-01
-3.83426160e-01 8.38304281e-01 8.99348021e-01 -7.72211015e-01
-1.18846345e+00 -8.25416923e-01 1.51298416e+00 -7.65610933e-01
3.08603853e-01 -5.58078110e-01 -9.26688313e-01 7.38329470e-01
6.89095259e-01 -7.32790828e-01 9.25362468e-01 3.39988649e-01
3.24337487e-03 5.46280146e-01 -1.19647980e+00 7.06915438e-01
1.36421692e+00 -4.69891876e-01 -9.66078639e-01 6.92007363e-01
7.47425616e-01 -7.84605265e-01 -7.50519812e-01 -2.00257659e-01
3.27123851e-01 -1.27621531e+00 4.95776623e-01 -4.69332546e-01
5.10719717e-01 -8.37691426e-02 -4.77400869e-02 -1.61631584e+00
-2.04239488e-01 -1.07526791e+00 1.17235690e-01 1.58361340e+00
3.27849358e-01 -8.60079110e-01 5.39591193e-01 1.34690988e+00
-3.66162926e-01 -3.10680121e-01 -6.98980868e-01 -8.63904320e-03
-2.48757407e-01 -3.49100828e-01 7.95007646e-01 1.25938845e+00
5.23655593e-01 3.63981366e-01 -4.99642402e-01 -9.44249034e-02
6.02680743e-01 4.65727821e-02 1.33069134e+00 -1.24834967e+00
-1.51878512e-02 -3.46998692e-01 3.75369102e-01 -1.16468620e+00
4.09853816e-01 -9.48165476e-01 1.39660344e-01 -1.97479451e+00
1.52922869e-01 -7.57782340e-01 4.70594883e-01 6.73610508e-01
-3.87929440e-01 -4.28804010e-02 2.79230177e-01 3.07062358e-01
-1.20579731e+00 4.57814544e-01 1.24433315e+00 -1.67728141e-02
-5.36032379e-01 -1.64210886e-01 -1.12933266e+00 6.93000436e-01
1.06653774e+00 -3.84846717e-01 -1.06077746e-01 -4.77144122e-01
3.83805394e-01 2.49667704e-01 3.13295752e-01 -8.12493205e-01
6.12592816e-01 -3.17268699e-01 -3.19465920e-02 -4.62323606e-01
2.77907014e-01 -1.12038232e-01 3.15343201e-01 3.77985299e-01
-9.08675492e-01 -2.45530844e-01 -3.16605657e-01 2.76599109e-01
-7.35417083e-02 -3.05383712e-01 2.48117656e-01 -4.16691840e-01
-5.69215715e-01 -1.75854042e-01 -8.52384448e-01 3.30701262e-01
1.03888869e+00 1.20915033e-01 -7.94659913e-01 -8.58773828e-01
-8.41014445e-01 8.98987770e-01 2.19486102e-01 4.41195041e-01
4.09090221e-01 -1.40474868e+00 -1.14821458e+00 -2.00410157e-01
1.19592622e-01 1.74833506e-01 6.72723830e-01 7.48681486e-01
-2.61512786e-01 6.34697556e-01 -4.21624556e-02 -3.09633076e-01
-1.49341571e+00 -1.22497432e-01 9.97747853e-02 -1.63243636e-01
-5.50182402e-01 7.88394749e-01 -7.97583982e-02 -9.18889344e-01
3.63353223e-01 -2.00963005e-01 -6.10302806e-01 4.12609071e-01
9.99706864e-01 4.80544627e-01 -3.46377432e-01 -4.15883124e-01
-1.70843989e-01 8.97944570e-02 8.44913349e-02 -1.41295895e-01
1.03344679e+00 -5.04734457e-01 -4.25506473e-01 4.93383408e-01
6.45223498e-01 1.67218447e-02 -8.17780554e-01 -6.44432008e-01
-6.56255893e-03 -5.79676032e-01 -5.92562854e-01 -9.68028605e-01
-2.90541351e-01 3.12380135e-01 -3.60412776e-01 7.75754035e-01
2.81719536e-01 1.51673242e-01 6.19232118e-01 8.25259924e-01
3.71112555e-01 -1.10261023e+00 5.23929060e-01 1.23674345e+00
1.42769372e+00 -1.32173789e+00 -3.43755245e-01 -7.31394708e-01
-9.36421216e-01 9.23143089e-01 7.30849862e-01 1.35561734e-01
1.77873820e-01 -1.82201773e-01 3.71991426e-01 -2.59561658e-01
-1.34506667e+00 1.35311913e-02 -1.54919565e-01 6.31668806e-01
5.92219770e-01 2.94093460e-01 -6.56762660e-01 1.23980737e+00
-8.72863948e-01 -2.40531832e-01 7.19541192e-01 6.39910698e-01
-4.94261861e-01 -9.20386493e-01 -4.14722115e-01 2.06307948e-01
9.93040763e-03 1.61515415e-01 -1.00831473e+00 2.40796179e-01
2.03014556e-02 1.63895106e+00 -2.09259195e-03 -4.49499667e-01
3.33457798e-01 4.25322801e-01 1.25363931e-01 -8.38219702e-01
-1.30325437e+00 -6.74048305e-01 9.92732406e-01 -5.01993060e-01
-5.62646747e-01 -7.30235398e-01 -1.07292211e+00 -8.11304390e-01
2.69428432e-01 5.82703590e-01 2.70629585e-01 1.21836901e+00
4.64920461e-01 3.00156415e-01 5.75960696e-01 -7.72919297e-01
-3.58683556e-01 -1.43660104e+00 1.32833242e-01 6.50208354e-01
7.08279908e-02 -4.23109084e-01 -3.80659282e-01 -2.03144625e-01]
|
[12.623285293579102, 8.037618637084961]
|
3d0649e9-2176-4f5e-9bbd-7376b4555239
|
demonstrating-the-feasibility-of-automatic
|
1603.03795
| null |
http://arxiv.org/abs/1603.03795v1
|
http://arxiv.org/pdf/1603.03795v1.pdf
|
Demonstrating the Feasibility of Automatic Game Balancing
|
Game balancing is an important part of the (computer) game design process, in
which designers adapt a game prototype so that the resulting gameplay is as
entertaining as possible. In industry, the evaluation of a game is often based
on costly playtests with human players. It suggests itself to automate this
process using surrogate models for the prediction of gameplay and outcome. In
this paper, the feasibility of automatic balancing using simulation- and
deck-based objectives is investigated for the card game top trumps.
Additionally, the necessity of a multi-objective approach is asserted by a
comparison with the only known (single-objective) method. We apply a
multi-objective evolutionary algorithm to obtain decks that optimise
objectives, e.g. win rate and average number of tricks, developed to express
the fairness and the excitement of a game of top trumps. The results are
compared with decks from published top trumps decks using simulation-based
objectives. The possibility to generate decks better or at least as good as
decks from published top trumps decks in terms of these objectives is
demonstrated. Our results indicate that automatic balancing with the presented
approach is feasible even for more complex games such as real-time strategy
games.
|
['Günter Rudolph', 'Boris Naujoks', 'Vanessa Volz']
|
2016-03-11
| null | null | null | null |
['real-time-strategy-games']
|
['playing-games']
|
[-2.13017583e-01 -6.27946779e-02 5.32418609e-01 4.85322252e-02
5.08264499e-03 -4.46170777e-01 4.14672345e-01 4.25703049e-01
-8.64717126e-01 9.77092505e-01 -4.22725022e-01 -4.42448616e-01
-6.78066373e-01 -1.00287294e+00 -3.43197674e-01 -5.90376616e-01
-1.57636274e-02 8.73809576e-01 2.52744555e-01 -7.46903360e-01
6.21580601e-01 5.98497808e-01 -2.11348009e+00 -3.42869945e-03
8.48403275e-01 7.34034121e-01 3.10831130e-01 9.26130116e-01
-1.03790045e-01 4.84399319e-01 -8.64589930e-01 -5.96781254e-01
2.75877893e-01 -7.67194986e-01 -4.12035644e-01 -1.85705498e-01
-6.20998442e-01 8.36783927e-03 5.59752941e-01 8.39530587e-01
5.87016582e-01 3.18155855e-01 6.65533185e-01 -1.19510555e+00
3.03479642e-01 4.89518732e-01 -2.67936379e-01 -1.09434240e-01
3.00324678e-01 1.68042973e-01 8.92728567e-01 -9.06692222e-02
3.82409364e-01 7.01418340e-01 3.25335413e-01 2.11045459e-01
-1.32683277e+00 -5.39745331e-01 -2.51967371e-01 2.11699516e-01
-1.40829182e+00 -1.05799742e-01 6.14533424e-01 -5.48650444e-01
7.56219447e-01 6.53558552e-01 1.43612099e+00 5.64489007e-01
5.19974887e-01 1.10139623e-01 1.24445343e+00 -6.88501060e-01
8.52566659e-01 4.82571512e-01 -3.81144643e-01 9.56581086e-02
7.11154044e-01 3.24944943e-01 -3.80249590e-01 6.89548776e-02
4.34928089e-01 -7.88115740e-01 -1.42666269e-02 -4.34949249e-01
-5.12030423e-01 9.28864002e-01 -6.67798892e-02 5.35729587e-01
-6.04487240e-01 6.21212684e-02 5.05290091e-01 1.75673619e-01
2.85100281e-01 1.13158357e+00 -1.42950937e-01 -8.42511714e-01
-1.09416902e+00 6.74792528e-01 1.16039455e+00 2.37598732e-01
4.29938108e-01 1.97041661e-01 1.47617996e-01 5.57195485e-01
2.56899029e-01 1.40528768e-01 5.46756089e-01 -5.65181792e-01
-1.56616226e-01 8.31948638e-01 1.63905576e-01 -1.07931292e+00
-4.74213839e-01 -5.91853499e-01 -4.19234395e-01 1.32957029e+00
3.97497743e-01 -1.27644077e-01 -2.15764672e-01 1.31747758e+00
3.11329722e-01 -3.40461224e-01 -8.17217305e-02 9.63120401e-01
3.19297999e-01 3.81541848e-01 -1.09135926e-01 -3.06549102e-01
1.11032844e+00 -3.09178233e-01 -5.46499491e-01 7.05560818e-02
5.06857991e-01 -8.53646219e-01 1.00884938e+00 8.03115308e-01
-1.14153016e+00 -4.71503228e-01 -1.08765686e+00 8.78797889e-01
-2.42186233e-01 -6.07838146e-02 4.57657456e-01 1.42035902e+00
-8.99230182e-01 8.02174509e-01 -4.81395423e-01 -1.71137437e-01
-2.72691011e-01 5.74634969e-01 -2.02332120e-02 7.10046113e-01
-1.03107500e+00 1.21579289e+00 7.82524526e-01 2.27650478e-01
-4.27841902e-01 -4.83897239e-01 -5.28698087e-01 2.45313630e-01
4.05680418e-01 -3.12509418e-01 9.62690532e-01 -1.10905159e+00
-2.11279273e+00 9.44792986e-01 4.83362108e-01 -3.52494299e-01
1.01063240e+00 3.79274011e-01 -8.53114128e-02 -3.08447748e-01
-1.02210924e-01 1.05111115e-01 3.96936655e-01 -1.13308644e+00
-6.26636147e-01 4.89671640e-02 3.44158322e-01 2.89931625e-01
-6.64053634e-02 6.26286566e-02 1.66980937e-01 -1.70690984e-01
-5.11844099e-01 -7.56142259e-01 -3.79714966e-01 -4.92874563e-01
-1.87926799e-01 1.97311074e-01 1.69670656e-02 -5.78622401e-01
1.42302716e+00 -1.54869306e+00 2.34254450e-01 5.65256596e-01
7.09986910e-02 2.76166707e-01 3.82819146e-01 4.01512504e-01
-6.83830157e-02 -1.13545336e-01 1.73600093e-01 2.36658799e-03
3.39780718e-01 -7.05160201e-02 3.56185794e-01 3.68984669e-01
-1.58420190e-01 4.21229422e-01 -6.22655332e-01 -3.95946354e-01
4.59876508e-01 -7.66071901e-02 -8.38989198e-01 2.43511885e-01
-1.29488617e-01 2.11669490e-01 -1.85311273e-01 1.54629558e-01
5.15680492e-01 3.42649311e-01 2.73615807e-01 2.06606269e-01
-6.40121281e-01 1.16900029e-02 -1.48056185e+00 9.74000335e-01
-6.33107841e-01 3.83813173e-01 -6.52814284e-02 -1.13516295e+00
1.39628994e+00 2.02536374e-01 5.28329015e-01 -8.05634260e-01
5.99739194e-01 5.61669052e-01 6.92435265e-01 -2.17389494e-01
9.54402983e-01 -5.51528454e-01 -1.12846687e-01 4.23418075e-01
-2.47902423e-01 -5.62471628e-01 6.29503548e-01 -3.27471077e-01
6.93704009e-01 2.66394466e-01 4.70176131e-01 -7.24332213e-01
6.62710905e-01 1.70540765e-01 1.09121993e-01 4.89583492e-01
1.48976594e-01 2.75529355e-01 8.83715689e-01 -2.30685651e-01
-1.22339654e+00 -6.12065613e-01 1.27122596e-01 7.24318743e-01
2.12482974e-01 -4.04799074e-01 -7.74390399e-01 1.15061000e-01
-2.61359632e-01 8.41037452e-01 -6.58059359e-01 -2.53426760e-01
-2.28053942e-01 -8.98017228e-01 2.53755122e-01 -1.00894175e-01
1.45788670e-01 -1.07970917e+00 -1.36567736e+00 6.07382953e-01
1.12409361e-01 -4.98151422e-01 3.29375267e-01 3.32112342e-01
-5.30328214e-01 -1.02670908e+00 -5.86605191e-01 -5.44988178e-02
2.44121060e-01 -3.48047674e-01 1.12913752e+00 3.69672090e-01
-4.32533085e-01 -6.35668635e-02 -6.32324815e-01 -5.88141859e-01
-8.19681883e-01 1.17250629e-01 -3.87002751e-02 -1.99097827e-01
3.19934748e-02 -7.05228448e-01 -3.46369624e-01 5.73757589e-01
-8.33376706e-01 2.55232006e-01 3.15762728e-01 5.53496063e-01
2.53860176e-01 4.06238616e-01 3.77822518e-01 -4.31452185e-01
9.68528807e-01 -4.81401309e-02 -1.13710451e+00 1.02463253e-01
-6.73937380e-01 1.72303811e-01 8.32740724e-01 -5.23756325e-01
-7.00700998e-01 -2.72705138e-01 -4.40379888e-01 5.43634780e-02
6.68425560e-02 4.65621322e-01 -1.77866757e-01 -3.06170136e-01
8.55155110e-01 1.00378178e-01 1.99736670e-01 -2.55789220e-01
-1.12861328e-01 4.73495811e-01 -2.46504750e-02 -9.12935734e-01
6.12385154e-01 -2.49601644e-03 1.52327523e-01 -7.29049921e-01
3.27093452e-01 -1.75713107e-01 -1.23653233e-01 -1.04700243e+00
5.87816000e-01 -2.35563859e-01 -1.44080484e+00 6.56480610e-01
-8.53922546e-01 -3.89066428e-01 -5.26826382e-01 6.26938164e-01
-8.82434726e-01 -8.87726694e-02 1.31250218e-01 -1.27553797e+00
-3.74367535e-02 -1.35872972e+00 5.13182282e-01 5.11745095e-01
-5.79884410e-01 -7.81396329e-01 3.39564800e-01 1.68784082e-01
3.58268470e-01 3.87419850e-01 7.14814067e-01 -4.98214811e-01
-2.53595680e-01 -4.25480843e-01 2.98876017e-01 1.03679866e-01
-5.02292588e-02 4.10430253e-01 -4.07293558e-01 -1.17759280e-01
-5.89594468e-02 -8.45316947e-02 6.09079003e-02 4.09484297e-01
5.71510255e-01 2.14727417e-01 7.32680336e-02 3.16083699e-01
1.69316196e+00 6.46143854e-01 9.21675324e-01 9.34331119e-01
-1.23796590e-01 7.40380228e-01 7.65647411e-01 8.91884923e-01
-8.89960900e-02 1.23873460e+00 6.29798353e-01 1.03975020e-01
3.58441323e-01 4.34326753e-03 2.50636578e-01 4.11025435e-01
-5.00119686e-01 -3.99638116e-01 -7.61984468e-01 4.67715144e-01
-1.67036402e+00 -9.87454057e-01 -3.90129268e-01 2.63948178e+00
3.58633965e-01 7.52594352e-01 6.73464060e-01 6.09262109e-01
1.00522768e+00 -2.29009688e-01 9.90526378e-02 -1.25150263e+00
8.34343955e-02 5.93489468e-01 4.56060112e-01 6.31456792e-01
-4.15580601e-01 4.23213363e-01 5.20575237e+00 1.03464866e+00
-1.13804650e+00 -1.78857833e-01 5.68617105e-01 -5.19502580e-01
-3.54462385e-01 1.38712421e-01 -4.17862594e-01 8.17794144e-01
7.95457900e-01 -8.18556845e-01 5.56488276e-01 5.70788860e-01
7.56694913e-01 -8.20190072e-01 -5.36478519e-01 7.22863495e-01
-2.65221387e-01 -1.21680605e+00 -3.43380570e-01 4.30986255e-01
3.58106643e-01 -8.03069174e-01 -2.54029542e-01 2.48818859e-01
1.01775847e-01 -1.16529548e+00 1.21846652e+00 3.56003284e-01
4.46795821e-01 -1.11522532e+00 9.41175699e-01 4.07636195e-01
-1.07839322e+00 1.62309155e-01 -1.38918966e-01 -5.79426467e-01
3.77487361e-01 5.84951520e-01 -6.68129385e-01 7.48747587e-01
4.78963256e-01 -2.50514090e-01 -3.35388422e-01 1.55218613e+00
4.78712060e-02 4.26154345e-01 -6.14156842e-01 -8.90284002e-01
2.35503450e-01 -6.86148286e-01 7.30538368e-01 8.06052446e-01
6.05914831e-01 -7.35452697e-02 -5.53510129e-01 1.07630944e+00
7.76740015e-01 6.97654486e-01 -3.18329722e-01 2.48248763e-02
4.81212363e-02 1.09549487e+00 -1.23158073e+00 1.96948752e-01
4.35196221e-01 6.41052663e-01 -1.64108872e-01 -2.15907887e-01
-1.09934437e+00 -4.92948741e-01 7.61670709e-01 4.14918333e-01
1.37047514e-01 8.90519470e-02 -6.42531097e-01 -5.66461980e-01
-6.94235638e-02 -8.87143731e-01 -1.10580131e-01 -6.62732303e-01
-5.59719563e-01 7.53418088e-01 1.47909805e-01 -1.31533730e+00
-5.00255942e-01 -6.74762130e-01 -8.31024766e-01 1.17604709e+00
-6.84729636e-01 -4.71074373e-01 -1.11469239e-01 -5.91199361e-02
1.13065891e-01 -3.05128485e-01 5.48437893e-01 2.01146200e-01
-3.31537485e-01 3.74033928e-01 3.03523958e-01 -6.16120398e-01
4.14609425e-02 -1.20157468e+00 -6.21736348e-02 6.22908235e-01
-1.85076147e-01 4.04157862e-03 1.41944945e+00 -5.34480810e-01
-7.60119438e-01 -4.39863093e-02 6.98618174e-01 1.16543315e-01
6.10898077e-01 -4.47120726e-01 -3.72100651e-01 -4.09799665e-01
1.97916925e-02 -7.88106680e-01 5.09941876e-01 2.28868350e-01
6.65964723e-01 -1.22703068e-01 -1.05067384e+00 7.50578284e-01
6.11309409e-01 7.83410221e-02 -3.21073204e-01 -2.90573508e-01
-2.23324642e-01 -3.99019510e-01 -5.12827635e-01 2.37742402e-02
8.50884795e-01 -1.59209096e+00 7.28396714e-01 -2.86940277e-01
4.58051294e-01 -2.23342389e-01 2.70841986e-01 -1.50059378e+00
-6.96040690e-02 -7.33070612e-01 6.75641894e-01 9.88696754e-01
3.84924352e-01 -5.60116708e-01 1.15155089e+00 6.45988345e-01
1.77305609e-01 -6.71910465e-01 -1.12674522e+00 -9.73204315e-01
4.87470217e-02 -3.96088004e-01 4.94248152e-01 4.74285811e-01
2.08192602e-01 -1.09351747e-01 -2.67210662e-01 -2.34553978e-01
4.53665286e-01 -5.30821644e-02 9.37346041e-01 -1.29595208e+00
-8.83759797e-01 -1.06364131e+00 -9.07956064e-01 -6.69241622e-02
-2.87133604e-02 -4.70929861e-01 -8.25988594e-03 -1.16819847e+00
-1.65531993e-01 -5.57618260e-01 -3.05886213e-02 -2.60475248e-01
-3.27299684e-02 1.35280088e-01 4.28521425e-01 -3.88881207e-01
-1.24037668e-01 3.22826892e-01 9.08838630e-01 2.20048353e-01
-5.24731576e-01 3.63671839e-01 -6.40416861e-01 4.46590722e-01
8.04809928e-01 -5.87996244e-01 -4.56778139e-01 4.37929839e-01
8.38371038e-01 3.04089952e-02 1.91744253e-01 -1.20858729e+00
-5.45093268e-02 -3.64294648e-01 -8.32757726e-02 -1.35832965e-01
2.52641022e-01 -8.67431045e-01 9.11935389e-01 8.69739354e-01
8.36665332e-02 1.40847666e-02 3.99212092e-01 5.21545485e-02
-1.31533235e-01 -8.63228321e-01 7.52122283e-01 -5.86402491e-02
-5.08020043e-01 -5.50879419e-01 -8.47335637e-01 -2.04935536e-01
1.33296418e+00 -9.95299041e-01 2.30446965e-01 -6.61757052e-01
-6.27690017e-01 -3.55263427e-02 5.97607315e-01 7.51820451e-04
8.83968174e-02 -8.82321537e-01 -7.88899362e-01 -1.41039133e-01
4.21416312e-02 -5.15734136e-01 3.93385440e-01 8.80113482e-01
-1.39845467e+00 7.74239674e-02 -8.20879936e-01 -1.51411667e-01
-1.46789491e+00 2.22066462e-01 6.84305310e-01 -6.41860068e-01
-1.99764505e-01 5.86879492e-01 -5.86086214e-02 -1.26865745e-01
-3.12679976e-01 -2.91641485e-02 -3.08874935e-01 2.04687700e-01
2.02073798e-01 5.89241445e-01 4.13449734e-01 -3.19414347e-01
-3.81564349e-01 3.07082593e-01 5.23655832e-01 -4.58388716e-01
1.33687937e+00 2.16001853e-01 -7.29942173e-02 2.25961402e-01
5.57946920e-01 2.77414471e-01 -9.53306794e-01 8.49123538e-01
-1.51267081e-01 -6.53835297e-01 -2.06839945e-02 -8.92946601e-01
-7.51683176e-01 4.36304450e-01 4.19550061e-01 4.67235804e-01
1.28683388e+00 -6.18847847e-01 7.02379867e-02 7.12437928e-02
7.78287053e-01 -1.21652842e+00 -2.15882689e-01 2.32476652e-01
7.28859246e-01 -4.83184576e-01 1.24457754e-01 -1.19523995e-01
-5.49199402e-01 1.16308939e+00 5.82304955e-01 -2.45282978e-01
2.79774368e-01 5.59456110e-01 -1.14748143e-01 3.38805467e-02
-5.85566580e-01 -4.66505468e-01 -5.16442321e-02 5.90097189e-01
2.40844160e-01 3.04966450e-01 -1.28122807e+00 8.27000201e-01
-6.32408917e-01 2.10138798e-01 9.22854602e-01 8.05511951e-01
-5.15862405e-01 -1.41758096e+00 -7.62801111e-01 4.01435882e-01
-1.39295101e-01 6.62015826e-02 -3.54177117e-01 1.14860296e+00
3.78117681e-01 9.51399267e-01 3.87304835e-02 -5.15647769e-01
5.83819389e-01 -2.00203612e-01 6.03155613e-01 -2.39474088e-01
-1.17459118e+00 -2.28984758e-01 4.26733822e-01 -1.88826025e-01
-8.98397416e-02 -3.85742545e-01 -7.52878428e-01 -9.43102598e-01
-6.26548171e-01 8.31513882e-01 9.74914610e-01 8.68810356e-01
-2.33145222e-01 6.47109330e-01 4.86216873e-01 -9.57129776e-01
-1.84509963e-01 -5.45438707e-01 -1.05939138e+00 1.24611348e-01
-5.62003613e-01 -9.31293011e-01 -3.59813839e-01 -4.20098454e-01]
|
[3.4809441566467285, 1.5828652381896973]
|
b0c10160-505f-4aa8-8cc3-c23801d807e9
|
source-free-adaptation-to-measurement-shift
|
2107.05446
| null |
https://arxiv.org/abs/2107.05446v3
|
https://arxiv.org/pdf/2107.05446v3.pdf
|
Source-Free Adaptation to Measurement Shift via Bottom-Up Feature Restoration
|
Source-free domain adaptation (SFDA) aims to adapt a model trained on labelled data in a source domain to unlabelled data in a target domain without access to the source-domain data during adaptation. Existing methods for SFDA leverage entropy-minimization techniques which: (i) apply only to classification; (ii) destroy model calibration; and (iii) rely on the source model achieving a good level of feature-space class-separation in the target domain. We address these issues for a particularly pervasive type of domain shift called measurement shift which can be resolved by restoring the source features rather than extracting new ones. In particular, we propose Feature Restoration (FR) wherein we: (i) store a lightweight and flexible approximation of the feature distribution under the source data; and (ii) adapt the feature-extractor such that the approximate feature distribution under the target data realigns with that saved on the source. We additionally propose a bottom-up training scheme which boosts performance, which we call Bottom-Up Feature Restoration (BUFR). On real and synthetic data, we demonstrate that BUFR outperforms existing SFDA methods in terms of accuracy, calibration, and data efficiency, while being less reliant on the performance of the source model in the target domain.
|
['Bernhard Schölkopf', 'Christopher K. I. Williams', 'Ian Mason', 'Cian Eastwood']
|
2021-07-12
|
source-free-adaptation-to-measurement-shift-1
|
https://openreview.net/forum?id=1JDiK_TbV4S
|
https://openreview.net/pdf?id=1JDiK_TbV4S
|
iclr-2022-4
|
['source-free-domain-adaptation']
|
['computer-vision']
|
[ 5.53899109e-01 -4.25018929e-03 -9.39306617e-02 -3.73679608e-01
-1.01302409e+00 -6.31760955e-01 6.12027526e-01 2.22186431e-01
-2.89474249e-01 8.24276805e-01 7.28825256e-02 4.54848334e-02
-9.06907097e-02 -6.56863689e-01 -8.67841661e-01 -7.95582592e-01
4.02870625e-01 4.40893799e-01 1.68165594e-01 -1.19151570e-01
1.33980080e-01 5.16477704e-01 -1.53225207e+00 6.51339293e-02
9.87093985e-01 1.23974705e+00 9.81054008e-02 5.01429319e-01
7.44541213e-02 4.22294706e-01 -6.99598372e-01 -2.23807737e-01
5.43379784e-01 -5.40844142e-01 -7.38611579e-01 4.16379213e-01
1.45479336e-01 -1.07780956e-01 -1.06213510e-01 9.27735806e-01
3.60741943e-01 1.80295840e-01 7.77030170e-01 -1.09110546e+00
-5.48406839e-01 -1.66622132e-01 -1.89452052e-01 -1.64791010e-02
3.74943197e-01 -8.00327659e-02 5.38126707e-01 -1.06079745e+00
7.10676312e-01 8.61997306e-01 7.14749575e-01 5.50042868e-01
-1.49575865e+00 -4.22312021e-01 2.62017190e-01 -3.56651619e-02
-1.36222827e+00 -8.48623097e-01 8.37366104e-01 -4.31419700e-01
7.40904450e-01 1.31462172e-01 3.92667770e-01 9.86349881e-01
1.44484982e-01 4.22472686e-01 9.75851178e-01 -6.75579011e-01
6.81381643e-01 4.34052587e-01 -4.15368341e-02 3.45507413e-01
1.03102051e-01 1.54378459e-01 -6.62843645e-01 -4.28382874e-01
3.98582876e-01 1.75364856e-02 -2.92457283e-01 -1.03216398e+00
-1.16561353e+00 7.83621848e-01 1.25421405e-01 7.78760910e-02
-6.34331584e-01 -5.05981982e-01 3.00703883e-01 6.45531416e-01
6.14149392e-01 3.56238902e-01 -8.80314052e-01 6.60632253e-02
-8.49309444e-01 3.25632483e-01 8.37043166e-01 1.03677118e+00
1.13835347e+00 3.19422260e-02 9.67035592e-02 8.87552798e-01
9.20614749e-02 6.02595806e-01 1.07701647e+00 -7.86932468e-01
3.81447583e-01 7.41590917e-01 3.62636894e-01 -5.71882844e-01
-9.68923494e-02 -5.11256695e-01 -4.97115314e-01 2.09804922e-01
4.17794734e-01 1.92144904e-02 -8.88953567e-01 1.92514634e+00
8.09192955e-01 2.19241362e-02 5.37356853e-01 5.03321648e-01
3.31314772e-01 4.19678330e-01 -2.20456064e-01 -2.82895029e-01
9.21665192e-01 -7.00798631e-01 -4.45562273e-01 -4.50982243e-01
6.13200426e-01 -6.01017296e-01 1.10188091e+00 3.57047737e-01
-7.94723630e-01 -7.49732912e-01 -1.26449931e+00 5.20529673e-02
-4.76654589e-01 1.28483698e-02 1.40247121e-01 6.79822981e-01
-8.00831497e-01 6.52250648e-01 -7.29395032e-01 -3.70857745e-01
3.24362695e-01 4.10845935e-01 -8.36610854e-01 -1.72172263e-01
-1.08464885e+00 9.05623257e-01 5.04899085e-01 -3.78641635e-01
-7.71143079e-01 -8.31627131e-01 -8.30687881e-01 -4.26688716e-02
3.18268180e-01 -5.47133267e-01 1.29801333e+00 -1.36972415e+00
-1.71328497e+00 6.34439945e-01 -3.20386916e-01 -3.96041751e-01
5.19877970e-01 -1.60913020e-01 -5.54969311e-01 7.92769045e-02
1.56270459e-01 2.96560913e-01 1.16695130e+00 -1.24261498e+00
-8.50737929e-01 -4.30607110e-01 -4.40020561e-01 2.70870268e-01
-5.36960602e-01 -4.41492707e-01 -7.54301473e-02 -5.59249580e-01
2.97715813e-01 -7.91252375e-01 6.96313530e-02 6.28185123e-02
-1.78983048e-01 1.59493938e-01 9.82384801e-01 -8.29499722e-01
1.12930357e+00 -2.31086493e+00 1.42365664e-01 3.45107496e-01
-3.73231322e-02 5.38259685e-01 -1.07340686e-01 3.91785055e-01
-2.31849596e-01 -3.89111280e-01 -5.94943345e-01 -3.27667743e-01
-7.24258050e-02 2.95937628e-01 -5.96173286e-01 5.36767364e-01
2.20676333e-01 5.49297094e-01 -8.35953057e-01 -2.05337390e-01
2.03267202e-01 3.74468386e-01 -7.60008395e-01 3.52450907e-01
2.09364314e-02 5.19849300e-01 -2.60406196e-01 5.03627419e-01
7.40619659e-01 -3.08643617e-02 2.27761775e-01 -8.28765109e-02
2.03909636e-01 3.78609329e-01 -1.37810016e+00 1.70317197e+00
-5.68159163e-01 1.85482144e-01 -1.42549083e-01 -1.05181336e+00
1.27604640e+00 3.26214582e-01 5.89438975e-01 -6.50016248e-01
-2.90411025e-01 4.23175901e-01 -3.33665550e-01 -1.84308052e-01
3.44803363e-01 -1.65033713e-01 -2.14240670e-01 3.31564903e-01
3.82686973e-01 1.08603062e-02 -1.66970968e-01 -8.51605460e-02
1.11606848e+00 3.56233120e-01 7.88786650e-01 -1.90631375e-01
7.02099741e-01 -2.02180147e-01 7.95464516e-01 5.70769310e-01
-1.94329038e-01 5.85133493e-01 1.82610363e-01 -3.37266058e-01
-1.06733537e+00 -1.25072038e+00 -2.04257801e-01 1.15451872e+00
-9.72367674e-02 -3.72425616e-02 -7.60014057e-01 -1.07107413e+00
1.61636025e-01 8.35210443e-01 -6.45303667e-01 -6.14150226e-01
-4.06800658e-01 -4.10162956e-01 2.77562678e-01 4.30070788e-01
5.49263656e-01 -6.62622690e-01 -5.92568219e-01 3.50990325e-01
-2.28136241e-01 -9.15700972e-01 -5.62896252e-01 5.31401277e-01
-9.40781176e-01 -9.68132257e-01 -4.38578159e-01 -6.44568026e-01
8.44233871e-01 5.99256530e-03 7.08167970e-01 -4.45474893e-01
3.74058127e-01 4.65363204e-01 -4.62904155e-01 -3.28809977e-01
-6.44306123e-01 2.42344931e-01 2.06495970e-01 4.01654541e-01
5.54978311e-01 -6.82096243e-01 -2.15522632e-01 3.50251555e-01
-9.74445641e-01 -3.44986916e-01 6.14494264e-01 1.05076265e+00
7.34921455e-01 -9.94344950e-02 9.43743408e-01 -9.97556090e-01
3.91848326e-01 -5.59651911e-01 -5.02146840e-01 2.30209142e-01
-1.02673435e+00 2.14079469e-01 8.01675677e-01 -7.28594840e-01
-1.14935672e+00 4.85747010e-01 8.16783383e-02 -5.34397960e-01
-2.95821130e-01 3.16559672e-01 -6.01007998e-01 -1.97459571e-02
1.10109997e+00 4.81730342e-01 6.05649687e-02 -7.39710450e-01
1.57892123e-01 9.03583705e-01 8.16050589e-01 -5.53870678e-01
8.62899303e-01 4.53496873e-01 -1.21475495e-01 -5.99039018e-01
-7.31312156e-01 -4.41114873e-01 -1.11661804e+00 2.88135856e-01
1.73057884e-01 -7.21726179e-01 2.57593989e-02 4.83846068e-01
-7.09739685e-01 -2.39364013e-01 -8.70313048e-01 3.90684545e-01
-6.52813852e-01 3.59758854e-01 1.73604265e-01 -5.81761360e-01
-2.22913459e-01 -8.35490406e-01 1.02515006e+00 5.69310300e-02
-1.45479232e-01 -1.18590093e+00 2.11898744e-01 -1.03964813e-01
3.49248141e-01 3.89957517e-01 8.34252536e-01 -1.15606523e+00
9.76487398e-02 -6.14140928e-01 1.86911374e-01 7.64349520e-01
4.94551897e-01 -4.95231777e-01 -1.28204203e+00 -5.64277649e-01
2.72810280e-01 -8.42167288e-02 5.63971221e-01 1.32183403e-01
7.22584188e-01 -4.23306376e-01 -2.72139847e-01 6.55319631e-01
1.11567688e+00 1.09918237e-01 3.82948190e-01 4.82772827e-01
3.45205128e-01 5.18444479e-01 6.73253298e-01 5.86010635e-01
4.30403918e-01 8.74087095e-01 3.95738631e-02 5.24484217e-02
-2.89144546e-01 -5.45126617e-01 5.08608103e-01 6.68089747e-01
4.18893129e-01 9.09946263e-02 -7.09076166e-01 7.80455351e-01
-1.83586502e+00 -5.70267797e-01 4.50759441e-01 2.65326858e+00
9.34844196e-01 1.39727518e-02 2.34434903e-01 3.28053027e-01
6.75172389e-01 -2.61287689e-01 -1.01351953e+00 -3.99394393e-01
-4.60777208e-02 2.24633783e-01 3.77693206e-01 4.73078132e-01
-1.11589408e+00 5.87373912e-01 5.65949488e+00 4.54035729e-01
-1.22394776e+00 1.72235489e-01 1.75882041e-01 3.82860675e-02
-1.15327515e-01 1.40665933e-01 -6.77369535e-01 5.24837434e-01
1.07786882e+00 -4.40954000e-01 4.62366343e-01 1.13738143e+00
-2.94926673e-01 6.74999431e-02 -1.44508278e+00 7.12011278e-01
4.46340963e-02 -8.02280784e-01 -2.19598766e-02 2.71474689e-01
5.37914336e-01 -2.67720312e-01 -8.85964110e-02 5.04385054e-01
1.94579333e-01 -4.96793836e-01 7.53071964e-01 4.89445359e-01
9.98152196e-01 -7.78999507e-01 6.82293952e-01 5.22957921e-01
-9.57544208e-01 -2.72629410e-01 -4.18990731e-01 1.56755418e-01
-2.37118021e-01 7.52034724e-01 -1.06316674e+00 7.82941043e-01
4.26449329e-01 5.36042333e-01 -6.22877955e-01 9.10396397e-01
-5.69605902e-02 5.41741014e-01 -2.73500830e-01 5.61106563e-01
-3.33905339e-01 5.45940883e-02 6.29501462e-01 9.66808498e-01
4.94838774e-01 -3.62264484e-01 1.71119034e-01 4.35749680e-01
1.46369301e-02 6.92800358e-02 -5.22556901e-01 1.47498518e-01
8.27745557e-01 8.41864526e-01 -2.63392836e-01 -2.02082649e-01
-3.07987869e-01 1.15686679e+00 2.34734535e-01 3.51967216e-01
-3.56450230e-01 -4.51957434e-01 7.63084412e-01 1.82213172e-01
6.15580380e-01 1.04602545e-01 -1.93248808e-01 -1.20727932e+00
2.60303825e-01 -8.20177317e-01 4.70421553e-01 -3.39475691e-01
-1.30266595e+00 6.14937782e-01 -5.57240397e-02 -1.50971437e+00
-6.80545568e-01 -2.39492267e-01 -1.01668173e-02 9.89375293e-01
-1.61478114e+00 -1.04636014e+00 -1.34665146e-01 8.29118907e-01
4.45486397e-01 -3.27579200e-01 1.17043912e+00 1.96279526e-01
-1.63958058e-01 9.08420324e-01 4.59019005e-01 -1.58369839e-01
9.44911599e-01 -1.12198341e+00 4.09174770e-01 8.14891577e-01
-5.61071672e-02 4.18107510e-01 6.04976296e-01 -5.81708610e-01
-1.13476038e+00 -1.32488489e+00 1.08578408e+00 -4.68033969e-01
2.66012669e-01 -4.08489883e-01 -1.26417875e+00 7.94460356e-01
-4.99055743e-01 2.44996905e-01 6.00215912e-01 -1.58339843e-01
-5.62312901e-01 -4.36832368e-01 -1.82149529e+00 5.46845682e-02
6.81896448e-01 -4.79383171e-01 -7.99081981e-01 1.24496043e-01
5.05954504e-01 -4.68670666e-01 -8.60587180e-01 2.92076081e-01
5.44693649e-01 -7.37534225e-01 9.25435901e-01 -5.88860393e-01
-9.17177498e-02 -3.59445602e-01 -5.33466995e-01 -1.60855448e+00
-3.35741729e-01 -4.63633746e-01 -3.60616475e-01 1.36426425e+00
1.20228082e-01 -1.13455403e+00 4.41356212e-01 6.97796524e-01
-6.55712187e-02 -4.45566684e-01 -1.19337380e+00 -1.01007342e+00
2.00569764e-01 -2.25335658e-01 9.82748568e-01 9.84943449e-01
-1.60520419e-01 1.55897409e-01 -2.00351015e-01 2.26100624e-01
4.67294574e-01 9.13672298e-02 9.19384718e-01 -1.46076560e+00
-3.46830159e-01 1.12489246e-01 -3.51148874e-01 -8.36782575e-01
1.22006916e-01 -8.00558269e-01 1.53485730e-01 -1.15478063e+00
-1.07170157e-01 -4.35973048e-01 -4.28718239e-01 6.71013415e-01
-2.07869574e-01 -8.75559524e-02 4.12161574e-02 5.86373389e-01
-3.01659852e-01 4.96892840e-01 7.54589498e-01 2.32793957e-01
-4.32541430e-01 1.47877440e-01 -8.83570910e-01 5.03451645e-01
6.61546409e-01 -6.94235623e-01 -3.81460100e-01 -4.05529663e-02
-2.50589311e-01 -6.13889210e-02 3.13538909e-01 -1.18113708e+00
6.25846535e-02 -1.49381936e-01 6.47979975e-01 -6.94905445e-02
3.22836190e-01 -9.86948788e-01 1.11440256e-01 3.86327386e-01
-3.28542024e-01 -1.75488487e-01 1.19375624e-01 6.52946472e-01
-1.30257726e-01 -2.76885778e-01 1.03224409e+00 2.73853779e-01
-6.11681879e-01 -6.35964945e-02 -1.65503830e-01 -3.77656855e-02
9.90137696e-01 -3.51240993e-01 -1.21336959e-01 -3.04377615e-01
-6.97808743e-01 -1.51621729e-01 7.71268785e-01 3.07443529e-01
3.53130817e-01 -1.42902088e+00 -6.01299942e-01 7.29499340e-01
3.16150486e-01 1.73179224e-01 2.94476915e-02 5.01850724e-01
1.66694835e-01 3.64318907e-01 -1.16979785e-01 -5.36249876e-01
-8.63893747e-01 7.51974344e-01 3.83359969e-01 -3.01133573e-01
-6.27881467e-01 5.65856516e-01 2.68464804e-01 -8.63543928e-01
-9.41119436e-03 -1.25587970e-01 1.44513562e-01 1.10515036e-01
4.58213300e-01 3.16231132e-01 6.22481763e-01 -7.32310951e-01
-4.02309775e-01 3.13294172e-01 -2.02634871e-01 -1.39171079e-01
1.43040705e+00 -3.72544408e-01 3.74543607e-01 3.70275587e-01
1.15893269e+00 8.54117388e-04 -1.77596509e+00 -8.05081367e-01
1.44447505e-01 -6.09591484e-01 2.66132131e-02 -1.06818962e+00
-6.55518353e-01 6.40541852e-01 8.10383379e-01 5.72273023e-02
1.51107347e+00 -2.28057936e-01 6.71194077e-01 3.25854957e-01
3.48869801e-01 -1.41695285e+00 -4.05212902e-02 2.41682440e-01
7.91106284e-01 -9.57271039e-01 4.15476263e-02 5.37124239e-02
-8.63776088e-01 1.09237611e+00 2.98361629e-01 -1.78467900e-01
6.00775719e-01 7.11621018e-03 -5.90643138e-02 2.09505051e-01
-6.61754191e-01 -8.14487487e-02 3.13527703e-01 1.09098494e+00
-1.62641361e-01 -1.39789522e-01 2.26543024e-01 8.54319751e-01
-1.76931277e-01 2.04311669e-01 2.62423396e-01 1.04433095e+00
-3.89249504e-01 -1.28933179e+00 -5.27887344e-01 3.54511648e-01
-5.84262051e-02 2.25012690e-01 -4.47538644e-01 7.25614190e-01
1.79922596e-01 8.82997990e-01 -1.08835980e-01 -2.81104058e-01
6.65580690e-01 6.51013076e-01 3.20792556e-01 -6.46643281e-01
-3.59639853e-01 9.22063664e-02 -8.35263580e-02 -4.07767594e-01
-1.22786179e-01 -1.05652714e+00 -9.99845147e-01 1.08857146e-02
-3.26894671e-01 1.95632204e-01 8.60548913e-01 1.10432410e+00
7.08627403e-01 3.03558767e-01 9.32268202e-01 -7.96678066e-01
-9.38539624e-01 -9.62508857e-01 -5.95812261e-01 5.32034755e-01
7.29792118e-01 -7.77147412e-01 -4.02012616e-01 3.95428538e-01]
|
[10.326803207397461, 3.171558380126953]
|
27efc540-619a-450e-8c17-9ada3e116b0c
|
universal-domain-adaptation
| null | null |
http://openaccess.thecvf.com/content_CVPR_2019/html/You_Universal_Domain_Adaptation_CVPR_2019_paper.html
|
http://openaccess.thecvf.com/content_CVPR_2019/papers/You_Universal_Domain_Adaptation_CVPR_2019_paper.pdf
|
Universal Domain Adaptation
|
Domain adaptation aims to transfer knowledge in the presence of the domain gap. Existing domain adaptation methods rely on rich prior knowledge about the relationship between the label sets of source and target domains, which greatly limits their application in the wild. This paper introduces Universal Domain Adaptation (UDA) that requires no prior knowledge on the label sets. For a given source label set and a target label set, they may contain a common label set and hold a private label set respectively, bringing up an additional category gap. UDA requires a model to either (1) classify the target sample correctly if it is associated with a label in the common label set, or (2) mark it as "unknown" otherwise. More importantly, a UDA model should work stably against a wide spectrum of commonness (the proportion of the common label set over the complete label set) so that it can handle real-world problems with unknown target label sets. To solve the universal domain adaptation problem, we propose Universal Adaptation Network (UAN). It quantifies sample-level transferability to discover the common label set and the label sets private to each domain, thereby promoting the adaptation in the automatically discovered common label set and recognizing the "unknown" samples successfully. A thorough evaluation shows that UAN outperforms the state of the art closed set, partial and open set domain adaptation methods in the novel UDA setting.
|
[' Michael I. Jordan', ' Jianmin Wang', ' Zhangjie Cao', ' Mingsheng Long', 'Kaichao You']
|
2019-06-01
| null | null | null |
cvpr-2019-6
|
['universal-domain-adaptation']
|
['computer-vision']
|
[ 4.04572606e-01 -3.25743929e-02 -5.72064757e-01 -4.32705015e-01
-6.15248144e-01 -8.39900494e-01 3.02137464e-01 1.13646790e-01
-1.98462978e-01 9.73475039e-01 -2.93774098e-01 3.17390710e-02
-7.31119663e-02 -9.81533349e-01 -6.79370880e-01 -8.18184614e-01
3.48530531e-01 8.34452331e-01 3.24411690e-01 -1.82287529e-01
-1.30161479e-01 2.74230838e-01 -1.48652244e+00 2.34454453e-01
9.56200123e-01 1.19258177e+00 2.41755828e-01 1.27639649e-02
-4.66037601e-01 4.00638610e-01 -7.91085184e-01 -3.36003423e-01
4.64200348e-01 -5.93667865e-01 -1.04264927e+00 1.79327969e-02
5.81086636e-01 -6.89377822e-03 9.68036428e-02 1.41416311e+00
3.35830152e-01 1.55550659e-01 1.02477300e+00 -1.41567445e+00
-8.64031434e-01 4.78613973e-01 -5.50404310e-01 1.44138232e-01
1.97965607e-01 -1.21676419e-02 1.02177703e+00 -5.64136624e-01
7.08130479e-01 9.51148629e-01 7.19405472e-01 1.02390969e+00
-1.36388063e+00 -1.13646555e+00 3.84644985e-01 3.97328921e-02
-1.59457266e+00 -1.60923436e-01 7.48513520e-01 -5.50714076e-01
3.85258824e-01 4.01306152e-02 1.02596104e-01 1.23696470e+00
-2.68984556e-01 4.97749209e-01 1.28776789e+00 -4.35236543e-01
3.82455885e-01 6.88719511e-01 3.74614596e-01 1.90574706e-01
8.45199004e-02 1.80724949e-01 -2.57470161e-01 -1.85585260e-01
5.91069043e-01 -6.79355562e-02 -2.26180568e-01 -6.90344930e-01
-1.06865144e+00 8.38128805e-01 3.60314906e-01 2.98113614e-01
-9.44819972e-02 -6.79541171e-01 4.70359087e-01 6.91189945e-01
4.77419794e-01 5.33408701e-01 -1.03106821e+00 3.58894527e-01
-4.49738801e-01 6.73514307e-02 8.29393685e-01 1.27075660e+00
1.39538205e+00 -3.34058851e-01 -7.71420598e-02 1.22431874e+00
1.65125489e-01 6.29163086e-01 8.77005935e-01 -6.70787334e-01
3.65472287e-01 1.09337890e+00 2.02152073e-01 -6.38587773e-01
-2.65041113e-01 -5.71463525e-01 -7.64349580e-01 2.24485829e-01
8.30885291e-01 -6.36532679e-02 -8.41951549e-01 2.36132193e+00
6.45675480e-01 4.17140603e-01 4.95645970e-01 6.04016781e-01
8.00923526e-01 4.16349232e-01 1.23398215e-01 -2.31150538e-01
1.38154018e+00 -6.79576695e-01 -4.37786192e-01 -4.07372683e-01
7.64530659e-01 -6.34641409e-01 1.24404621e+00 -5.82218841e-02
-3.98843557e-01 -7.78766394e-01 -1.15603387e+00 2.34283775e-01
-7.16152847e-01 -1.02948677e-02 2.06205934e-01 7.37607956e-01
-4.81491596e-01 3.74543518e-01 -5.45693599e-02 -6.43931031e-01
4.74293441e-01 3.64740819e-01 -5.30907512e-01 -3.43401700e-01
-1.39574289e+00 7.21418798e-01 8.57801080e-01 -6.35135531e-01
-7.41419911e-01 -8.70407879e-01 -7.43678749e-01 -5.46029815e-03
6.20971084e-01 -6.16891205e-01 1.24090648e+00 -1.53558147e+00
-1.26105034e+00 1.25126374e+00 2.72982307e-02 -2.82491446e-01
3.65334153e-01 1.27280727e-01 -8.58865857e-01 -2.52521753e-01
4.43289936e-01 5.72481453e-01 7.97536612e-01 -1.32850742e+00
-1.06611550e+00 -5.11248529e-01 5.18532991e-02 2.09692016e-01
-5.17026842e-01 -3.55903447e-01 2.76259203e-02 -6.69594049e-01
9.28722098e-02 -8.70069861e-01 2.95092911e-01 -1.74520358e-01
-1.41173050e-01 -5.30295074e-01 1.07835293e+00 -2.65996814e-01
1.16480172e+00 -2.29989409e+00 -1.40764356e-01 2.80897051e-01
7.95720890e-02 4.07820255e-01 -3.14833134e-01 -2.30428632e-02
-2.80635178e-01 -1.38481602e-01 -3.98043633e-01 1.43480062e-01
-1.14452802e-01 5.64378858e-01 -3.37180585e-01 2.46592566e-01
6.12948313e-02 4.20464993e-01 -1.07031620e+00 -3.27458501e-01
-1.87285602e-01 -1.44642651e-01 -3.56688678e-01 3.86033744e-01
-3.24234873e-01 5.98420978e-01 -3.69342625e-01 6.58762872e-01
7.98646748e-01 -3.99030954e-01 2.82292068e-01 -1.55143157e-01
4.33014154e-01 -1.08242445e-01 -1.57071090e+00 1.29153264e+00
-3.19095641e-01 2.21325412e-01 -4.09760848e-02 -1.03494465e+00
1.41743958e+00 2.45605737e-01 5.01317739e-01 -5.80172181e-01
5.89406751e-02 5.07019877e-01 8.56312085e-03 -2.88705498e-01
1.97222844e-01 -5.87226987e-01 -2.13353738e-01 5.54126740e-01
4.32564229e-01 3.66329372e-01 -5.38001694e-02 -6.12200163e-02
8.22091222e-01 -2.95356773e-02 5.99977672e-01 -4.09190685e-01
7.75833666e-01 -1.44989789e-01 1.01760840e+00 7.53809631e-01
-6.28448486e-01 5.01503885e-01 2.29957208e-01 -3.17180127e-01
-9.82443273e-01 -1.10566390e+00 -3.37513477e-01 1.51312280e+00
3.84562641e-01 1.81558162e-01 -5.23638487e-01 -1.30858231e+00
1.75342813e-01 5.40349483e-01 -8.03035975e-01 -5.78250885e-01
-2.46114135e-01 -3.49283278e-01 4.28172112e-01 4.39491093e-01
6.56265676e-01 -1.00271869e+00 -1.82086357e-03 1.53596476e-01
-3.87878954e-01 -9.84255373e-01 -7.14590192e-01 3.91900450e-01
-6.45389318e-01 -1.33744979e+00 -6.96010351e-01 -1.10935307e+00
6.93832040e-01 1.71495989e-01 1.14863622e+00 -4.05509621e-01
3.33478212e-01 2.28897914e-01 -4.09940749e-01 -3.30490351e-01
-7.82635689e-01 1.79161876e-01 2.95259833e-01 4.42194343e-01
9.18739080e-01 -4.30612475e-01 -1.67615280e-01 8.40078652e-01
-6.84422672e-01 -3.34649831e-01 2.77843952e-01 9.20486093e-01
6.67468309e-01 2.80084252e-01 1.17993176e+00 -1.29350138e+00
4.91548061e-01 -9.72135246e-01 -3.95049483e-01 5.19430339e-01
-7.42074013e-01 -1.54592460e-02 7.32987463e-01 -9.87272441e-01
-1.03199077e+00 1.09437279e-01 2.36969709e-01 -4.20607656e-01
-5.63848615e-01 3.11384529e-01 -7.14745581e-01 1.70504391e-01
1.10221493e+00 1.19509414e-01 9.72371772e-02 -6.71548665e-01
2.31481761e-01 8.06753278e-01 6.71849608e-01 -8.01435411e-01
6.55184984e-01 2.89328814e-01 -2.14511320e-01 -3.63317907e-01
-1.10325229e+00 -8.67035151e-01 -9.71857846e-01 8.62080306e-02
5.98128974e-01 -9.13839757e-01 -1.14974514e-01 4.86013949e-01
-8.44169021e-01 -2.77218044e-01 -7.01278806e-01 3.30623657e-01
-4.21911806e-01 3.45554680e-01 8.04926604e-02 -3.65637034e-01
-7.78321922e-02 -9.78005588e-01 6.54077411e-01 3.26432228e-01
-3.51682663e-01 -1.17466748e+00 9.41223949e-02 9.50418264e-02
1.13287792e-01 8.99928659e-02 1.15311742e+00 -1.29971409e+00
-4.38146759e-03 -2.79355079e-01 -3.02881598e-01 7.17139721e-01
5.54881573e-01 -5.45945883e-01 -9.98148978e-01 -5.21219969e-01
-2.02179998e-01 -3.62057269e-01 5.28999805e-01 1.37531519e-01
7.40693808e-01 -2.62728900e-01 -4.49236691e-01 4.71251220e-01
1.33429193e+00 3.46411109e-01 2.53266603e-01 3.67993623e-01
5.09522259e-01 5.78035474e-01 7.13945806e-01 3.28915447e-01
2.53200889e-01 7.57814288e-01 1.51144147e-01 3.25370282e-02
-2.82375515e-01 -3.69110823e-01 3.28206182e-01 5.28686166e-01
4.36057627e-01 -2.33507559e-01 -8.74764740e-01 7.76331306e-01
-1.75373149e+00 -6.61376178e-01 1.84132963e-01 2.38050246e+00
1.22714627e+00 -5.07666729e-02 3.45996261e-01 -4.39901128e-02
1.19175816e+00 -3.84186625e-01 -1.03140843e+00 -2.12091699e-01
-3.44208390e-01 1.13044698e-02 5.05888045e-01 2.73586452e-01
-1.31909549e+00 8.45763445e-01 6.09317780e+00 1.09366131e+00
-9.03556764e-01 2.14898363e-01 3.84057462e-01 4.41202343e-01
-1.23049207e-01 -6.70945942e-02 -1.12605381e+00 6.79615140e-01
7.95147896e-01 -3.30645710e-01 1.85533658e-01 9.91494358e-01
-4.67471659e-01 2.24750534e-01 -1.39125967e+00 8.00151527e-01
-1.09611444e-01 -8.60752106e-01 2.33539492e-01 1.26954317e-01
1.01990306e+00 -1.25410572e-01 6.76407292e-02 7.75278449e-01
6.73628926e-01 -6.51680052e-01 3.60859066e-01 2.06548765e-01
1.13313043e+00 -6.58106983e-01 9.23121393e-01 6.03178442e-01
-1.06480885e+00 -2.81365693e-01 -5.35765231e-01 1.36053354e-01
-4.14450169e-01 4.12360698e-01 -9.80108023e-01 4.91440028e-01
6.94561303e-01 7.83468246e-01 -5.60365796e-01 8.20895493e-01
-2.70816963e-02 5.61681032e-01 -3.69548500e-01 4.61182892e-01
2.06839759e-02 -1.49956793e-01 4.89094257e-01 9.72883463e-01
2.75676727e-01 -6.23208173e-02 4.92640495e-01 7.57000566e-01
-4.32715505e-01 2.20696270e-01 -6.56010747e-01 1.89286962e-01
8.21723282e-01 8.11260223e-01 -4.45816636e-01 -4.40490186e-01
-5.30962825e-01 8.98393095e-01 3.25146675e-01 4.65151995e-01
-5.90592325e-01 -2.60784656e-01 9.34898555e-01 1.30043253e-02
1.69104919e-01 5.08001685e-01 -3.32302511e-01 -1.12497962e+00
-9.14396644e-02 -7.87317634e-01 1.06529403e+00 -4.90147531e-01
-1.99047589e+00 3.61641675e-01 -3.69917378e-02 -1.55143094e+00
-1.76763624e-01 -5.42085767e-01 -1.99498042e-01 1.04611003e+00
-1.50218976e+00 -1.16705990e+00 -3.47606629e-01 1.02218735e+00
3.43478978e-01 -7.40283728e-01 1.01371050e+00 5.07125616e-01
-3.20301414e-01 1.20043588e+00 4.86969113e-01 2.45774359e-01
1.22167265e+00 -1.22198927e+00 -1.67590380e-01 4.86653507e-01
-2.21310511e-01 5.02595901e-01 4.50612426e-01 -5.53649902e-01
-5.22342265e-01 -1.45344293e+00 7.09387183e-01 -5.42080998e-01
5.18790543e-01 -1.70414388e-01 -1.45619440e+00 9.23749864e-01
-3.33487004e-01 1.45689517e-01 9.95720923e-01 4.20725733e-01
-9.92923200e-01 -2.97156304e-01 -1.55584991e+00 1.22654840e-01
8.35688889e-01 -5.84341824e-01 -6.39821470e-01 2.36738190e-01
8.09348702e-01 -1.06614985e-01 -1.02582717e+00 5.42244434e-01
4.75545526e-01 -5.33670008e-01 9.60074008e-01 -8.42651188e-01
1.97961211e-01 -4.92264450e-01 -3.59453559e-01 -1.45818520e+00
-5.17858028e-01 2.72300150e-02 -1.14745766e-01 1.64719951e+00
3.14338565e-01 -8.70810628e-01 6.66838169e-01 3.31568867e-01
1.04793161e-01 -2.54263610e-01 -1.16339421e+00 -1.35591328e+00
5.29960990e-01 -4.11723964e-02 1.10549438e+00 1.66509664e+00
-3.00291836e-01 5.22576630e-01 -2.85070121e-01 2.03321114e-01
4.12040591e-01 1.87959999e-01 7.72742689e-01 -1.79771745e+00
-1.19057953e-01 -4.41381812e-01 -3.18485618e-01 -9.43982124e-01
3.71182829e-01 -1.20987785e+00 -1.57997325e-01 -8.98813844e-01
3.75111639e-01 -9.44194078e-01 -7.31523216e-01 6.94950044e-01
-1.60643950e-01 1.70206875e-01 -8.89075547e-02 5.14005840e-01
-6.21220648e-01 4.27437127e-01 1.18923318e+00 -2.80043274e-01
-3.61750096e-01 1.75972939e-01 -1.03171551e+00 6.77240074e-01
8.34010601e-01 -5.59272826e-01 -4.58814114e-01 -1.24422610e-01
-3.66857350e-02 -3.29037130e-01 1.70313656e-01 -1.00525427e+00
-1.51335504e-02 -4.04907316e-01 1.67267308e-01 -2.05459490e-01
-6.09461963e-02 -1.20151091e+00 8.37424491e-03 2.11605191e-01
-4.75303471e-01 -4.50306445e-01 1.75124168e-01 7.24520564e-01
-1.63562834e-01 -4.57977235e-01 1.35227621e+00 -2.32458729e-02
-1.16277170e+00 3.02300364e-01 6.66531995e-02 4.02332902e-01
1.32179809e+00 -4.48431343e-01 -3.44501674e-01 -3.96913812e-02
-7.84810722e-01 2.19865397e-01 5.62357068e-01 5.05902469e-01
1.88630491e-01 -1.57175052e+00 -7.89535165e-01 5.36594331e-01
7.85075665e-01 6.35977089e-03 2.57884800e-01 1.88464224e-01
1.88409269e-01 2.28492133e-02 -3.56082231e-01 -6.75775707e-01
-1.19749546e+00 8.21448267e-01 5.04928172e-01 -3.53675455e-01
-3.57710302e-01 8.04096222e-01 7.26542175e-01 -1.13074505e+00
1.23727180e-01 8.87942761e-02 -4.02579039e-01 1.55823454e-01
5.16285837e-01 2.97217369e-01 -1.97779506e-01 -8.92409384e-01
-2.30295748e-01 5.56462646e-01 -2.20549196e-01 5.19485056e-01
7.80509472e-01 -3.08727771e-01 3.96115407e-02 5.19795954e-01
1.36685038e+00 -3.27017665e-01 -1.17880321e+00 -1.13353229e+00
-2.77903862e-03 -3.34160358e-01 -4.23931688e-01 -1.28891516e+00
-8.64245772e-01 6.31043792e-01 1.02541578e+00 4.88900915e-02
1.17566490e+00 2.10036933e-01 6.88030005e-01 2.23243624e-01
3.39263380e-01 -1.27383494e+00 1.17121413e-01 7.04727590e-01
5.16178727e-01 -1.57038057e+00 -1.91298917e-01 -4.28430229e-01
-6.96406007e-01 8.54272366e-01 1.04784310e+00 2.00667500e-01
6.27260923e-01 -1.44396067e-01 2.15154439e-01 9.56776366e-02
-3.39343876e-01 -1.96708247e-01 3.29035699e-01 1.12268376e+00
2.12209523e-02 4.35266495e-01 -2.59622652e-02 1.00264812e+00
-3.23887840e-02 -7.99128711e-02 2.06610069e-01 5.73744953e-01
-6.14981413e-01 -1.23114729e+00 -4.94679183e-01 3.23330820e-01
-1.32486939e-01 2.63279259e-01 -5.21111548e-01 7.82757342e-01
6.95963085e-01 7.80685544e-01 7.02008139e-03 -4.57382917e-01
5.31724036e-01 4.70646799e-01 2.41175070e-01 -8.18108618e-01
-4.07450229e-01 -2.84856230e-01 -2.85633624e-01 2.25038640e-03
-3.40229630e-01 -5.62882543e-01 -1.14729273e+00 -1.56686604e-01
-3.35898697e-01 7.88655654e-02 1.04057528e-01 9.56574261e-01
3.11969191e-01 2.36124545e-01 5.18225610e-01 -1.09328106e-02
-7.64515102e-01 -9.99526858e-01 -9.13767874e-01 9.74131107e-01
3.76742095e-01 -1.04069197e+00 -2.70246387e-01 2.10255161e-01]
|
[10.358174324035645, 3.158475160598755]
|
147e217d-ccbd-47b7-bc65-0ace969913c4
|
corenet-coherent-3d-scene-reconstruction-from
|
2004.12989
| null |
https://arxiv.org/abs/2004.12989v2
|
https://arxiv.org/pdf/2004.12989v2.pdf
|
CoReNet: Coherent 3D scene reconstruction from a single RGB image
|
Advances in deep learning techniques have allowed recent work to reconstruct the shape of a single object given only one RBG image as input. Building on common encoder-decoder architectures for this task, we propose three extensions: (1) ray-traced skip connections that propagate local 2D information to the output 3D volume in a physically correct manner; (2) a hybrid 3D volume representation that enables building translation equivariant models, while at the same time encoding fine object details without an excessive memory footprint; (3) a reconstruction loss tailored to capture overall object geometry. Furthermore, we adapt our model to address the harder task of reconstructing multiple objects from a single image. We reconstruct all objects jointly in one pass, producing a coherent reconstruction, where all objects live in a single consistent 3D coordinate frame relative to the camera and they do not intersect in 3D space. We also handle occlusions and resolve them by hallucinating the missing object parts in the 3D volume. We validate the impact of our contributions experimentally both on synthetic data from ShapeNet as well as real images from Pix3D. Our method improves over the state-of-the-art single-object methods on both datasets. Finally, we evaluate performance quantitatively on multiple object reconstruction with synthetic scenes assembled from ShapeNet objects.
|
['Pablo Bauszat', 'Vittorio Ferrari', 'Stefan Popov']
|
2020-04-27
| null |
https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/3439_ECCV_2020_paper.php
|
https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123470358.pdf
|
eccv-2020-8
|
['3d-scene-reconstruction']
|
['computer-vision']
|
[ 2.65023224e-02 2.74967372e-01 3.48925859e-01 -3.01330030e-01
-8.98285687e-01 -6.04438663e-01 7.29343593e-01 -1.32577851e-01
-1.34691015e-01 6.58768475e-01 2.88772583e-01 1.25575513e-01
1.00920193e-01 -9.49318409e-01 -1.33789814e+00 -4.67350602e-01
2.45328799e-01 1.14762533e+00 4.01070505e-01 1.26619518e-01
-2.48505585e-02 1.14169860e+00 -1.30449975e+00 4.34618771e-01
2.38861263e-01 1.00155187e+00 4.26279455e-01 7.02588260e-01
-1.15178026e-01 7.29300559e-01 -3.13654840e-01 -3.65755796e-01
6.52631044e-01 -1.50375754e-01 -8.02140474e-01 5.61157465e-01
1.06396914e+00 -8.11833620e-01 -6.10795617e-01 5.22626877e-01
2.56085336e-01 -1.53265819e-01 6.16309762e-01 -6.31646514e-01
-8.22963655e-01 1.24710932e-01 -6.50557220e-01 -3.42300028e-01
1.79860085e-01 3.05911869e-01 9.41538990e-01 -1.21512520e+00
1.09649396e+00 1.53351629e+00 7.27348566e-01 3.68728757e-01
-1.78810573e+00 -1.98034376e-01 5.03876656e-02 -2.29210719e-01
-1.26198256e+00 -4.61126894e-01 7.73870826e-01 -3.92374963e-01
1.33570135e+00 1.35067269e-01 8.52156579e-01 9.33112025e-01
2.18241721e-01 5.85960209e-01 6.84550464e-01 -1.01921260e-01
9.90451593e-03 -3.52957070e-01 -3.01243097e-01 8.70426238e-01
2.82212019e-01 4.40124350e-05 -3.87045085e-01 -2.73937546e-02
1.47716212e+00 9.91898030e-02 -1.70521170e-01 -9.15071249e-01
-1.47431087e+00 6.15595698e-01 5.95221162e-01 -1.17780574e-01
-4.10427451e-01 7.41480112e-01 1.02178296e-02 6.10634014e-02
5.46820045e-01 3.27590555e-01 -4.61158067e-01 3.51558924e-01
-7.53847301e-01 8.02410901e-01 8.51930618e-01 1.12636375e+00
9.14769590e-01 2.53899395e-01 1.38442090e-03 4.99429286e-01
4.81758386e-01 6.95244730e-01 -1.53465316e-01 -1.32673848e+00
4.90450799e-01 4.34636474e-01 2.69902915e-01 -8.01047325e-01
-4.50649232e-01 -5.53698301e-01 -6.72593713e-01 5.42107522e-01
3.36980581e-01 4.16627437e-01 -1.05181921e+00 1.64993966e+00
3.10544193e-01 2.34403878e-01 -1.01668142e-01 1.03555048e+00
1.06897974e+00 6.72651529e-01 -3.27155769e-01 1.89545617e-01
1.19371140e+00 -9.24437642e-01 -2.77172267e-01 -2.01886937e-01
2.06525743e-01 -7.96445370e-01 7.90900052e-01 3.05288196e-01
-1.63497782e+00 -4.90443021e-01 -9.66386199e-01 -6.30028784e-01
9.19356346e-02 2.99971667e-04 4.84680265e-01 2.82731466e-03
-1.11485946e+00 6.32987916e-01 -9.96913075e-01 7.28966817e-02
7.79174864e-01 3.06582510e-01 -5.51273584e-01 -1.95789889e-01
-4.55005616e-01 8.94166768e-01 1.37193158e-01 -1.52925670e-01
-1.28386867e+00 -1.08413827e+00 -8.82764995e-01 -1.02479709e-02
2.86868453e-01 -1.28325760e+00 1.30207324e+00 -7.05681562e-01
-1.22288191e+00 1.14169407e+00 2.78635370e-03 -4.22452211e-01
7.20395327e-01 -1.68340698e-01 2.53972083e-01 1.81346208e-01
1.90290306e-02 9.16249156e-01 6.08599305e-01 -1.82209516e+00
-9.92547255e-03 -5.92249215e-01 1.40866429e-01 3.73044312e-01
5.90890110e-01 -4.83939350e-01 -5.00643432e-01 -6.53744400e-01
5.22982776e-01 -7.60618150e-01 -1.39444962e-01 7.50484824e-01
-4.58544314e-01 2.58014917e-01 9.96048748e-01 -6.97089195e-01
1.74646035e-01 -1.83835542e+00 6.48432195e-01 -1.23297535e-01
4.85185653e-01 -7.29308929e-03 -3.66829127e-01 2.12326124e-01
6.33064061e-02 -8.06354880e-02 -4.48985040e-01 -1.05917752e+00
-4.67754304e-02 5.90571523e-01 -5.36378026e-01 6.10343039e-01
3.00466001e-01 1.39198673e+00 -6.22899413e-01 -2.66227275e-01
3.94593656e-01 8.36263835e-01 -8.28687191e-01 3.43270451e-01
-6.48169994e-01 7.04350293e-01 -2.81427264e-01 6.05082989e-01
8.15720022e-01 -4.25008863e-01 -1.21594690e-01 -5.07187545e-01
-5.57916276e-02 4.12024200e-01 -1.23179853e+00 2.21018791e+00
-5.33426464e-01 4.26817656e-01 3.77795368e-01 -9.12842870e-01
8.01550865e-01 2.82263041e-01 6.36238754e-01 -5.46998560e-01
-4.15083580e-02 8.01474303e-02 -3.33563477e-01 -2.72831112e-01
4.25059348e-01 -3.81324649e-01 1.92248195e-01 5.09378850e-01
3.65382969e-01 -7.96178937e-01 -4.11866277e-01 1.19478211e-01
9.23754871e-01 5.58443606e-01 -2.57871468e-02 -9.44530740e-02
3.04228207e-03 -1.62310600e-01 2.54667401e-01 5.66696048e-01
3.86333048e-01 1.45493388e+00 2.13481128e-01 -6.93738043e-01
-1.56219745e+00 -1.59964907e+00 -1.89809635e-01 3.14864129e-01
2.82744825e-01 -2.33777300e-01 -3.51683736e-01 -3.57885361e-01
2.67083734e-01 6.36837423e-01 -5.80868661e-01 1.26873076e-01
-9.22613442e-01 -3.35231692e-01 3.42748821e-01 5.25977254e-01
3.98243457e-01 -1.12140286e+00 -8.51894081e-01 3.00939381e-01
-2.13395253e-01 -1.15270126e+00 -3.10945272e-01 1.16592251e-01
-9.67338502e-01 -9.86300886e-01 -6.13178134e-01 -5.97512066e-01
6.70558631e-01 3.86359930e-01 1.48754930e+00 1.31336138e-01
-5.10813415e-01 5.35081387e-01 1.60078615e-01 -8.18155929e-02
-4.48625088e-01 -3.24207842e-01 -1.78847179e-01 -4.96312156e-02
-3.69425088e-01 -9.00977254e-01 -4.40369219e-01 1.02710590e-01
-9.89053905e-01 5.66718459e-01 3.88878644e-01 5.85057974e-01
1.06273222e+00 -4.21003610e-01 2.63948500e-01 -6.71162546e-01
-7.69606382e-02 -3.62272859e-01 -6.66717827e-01 1.00596435e-01
3.68800126e-02 1.96362793e-01 4.24219966e-01 -3.10230076e-01
-1.08853757e+00 2.92610139e-01 -2.40147352e-01 -8.29629719e-01
-9.01169330e-02 -6.68867081e-02 -2.26764932e-01 -8.74331295e-02
5.10437965e-01 3.17732006e-01 7.09848255e-02 -8.06146860e-01
5.51353335e-01 -9.08069685e-02 5.90224802e-01 -7.08366394e-01
7.39000142e-01 8.65379870e-01 3.36492896e-01 -7.31977880e-01
-8.06212306e-01 -5.00642769e-02 -9.92514789e-01 6.61147060e-03
9.04895306e-01 -1.01584411e+00 -7.23225176e-01 4.39647734e-01
-1.64535308e+00 -4.16263491e-01 -7.02337086e-01 3.31675351e-01
-1.02637517e+00 2.12471977e-01 -6.81324303e-01 -3.61054182e-01
-1.47464320e-01 -1.35408020e+00 1.60569298e+00 -2.84740955e-01
-3.95089164e-02 -7.93889165e-01 9.93704516e-03 3.01103085e-01
2.31711715e-01 5.22054911e-01 9.16084945e-01 3.66025381e-02
-1.41246307e+00 1.80776492e-01 -4.40312475e-01 1.81642413e-01
-9.74225998e-02 -2.55540222e-01 -9.26457405e-01 -1.91388190e-01
1.21267892e-01 -5.08270979e-01 9.37713563e-01 4.43398297e-01
1.22362769e+00 -4.35593069e-01 -1.84481576e-01 9.94575500e-01
1.60665548e+00 -2.77659684e-01 6.77248478e-01 -5.22120409e-02
1.11559355e+00 5.01668453e-01 -1.34047821e-01 2.78205901e-01
5.52245855e-01 9.87493277e-01 9.42173958e-01 -1.14718162e-01
-8.44530642e-01 -4.64902788e-01 -4.01554368e-02 5.80317855e-01
-1.84673965e-01 -2.45374158e-01 -7.74727762e-01 5.26031494e-01
-1.71502805e+00 -8.06664824e-01 -2.57545084e-01 2.11606026e+00
7.51234710e-01 -7.19078556e-02 -1.89692184e-01 -3.19259733e-01
1.41488519e-02 3.01414073e-01 -6.97949231e-01 -2.17795800e-02
-3.67398083e-01 2.79824793e-01 4.24941361e-01 8.16766202e-01
-7.02621639e-01 8.11522961e-01 6.38781071e+00 4.11634475e-01
-1.05137682e+00 2.06171185e-01 4.52896804e-01 -4.75471795e-01
-8.26412141e-01 9.21975970e-02 -6.17823899e-01 -3.23145203e-02
3.39730501e-01 2.40127012e-01 4.59696352e-01 5.25711119e-01
-1.15096375e-01 5.65151982e-02 -1.35750723e+00 1.11767983e+00
2.73872524e-01 -1.80148387e+00 4.12629157e-01 2.28624925e-01
8.03113937e-01 3.18974286e-01 -1.48665994e-01 -1.76910460e-01
1.99933901e-01 -1.15888047e+00 1.33781528e+00 7.54833937e-01
9.78978753e-01 -3.83590996e-01 8.07654485e-02 4.47287500e-01
-1.05210900e+00 3.43576789e-01 -4.50678349e-01 2.22579706e-02
4.72390771e-01 5.66919684e-01 -8.31406832e-01 6.58114254e-01
5.54311574e-01 7.44200587e-01 -3.37697119e-01 9.16841209e-01
-1.54634183e-02 2.07561851e-01 -4.65113610e-01 5.76544106e-01
1.37940645e-01 -1.98619366e-01 8.87673080e-01 8.00905406e-01
3.32463354e-01 2.42574960e-01 5.62777109e-02 1.59725094e+00
-3.01012307e-01 -3.52034926e-01 -7.59372473e-01 4.17467892e-01
2.11917758e-01 8.43948364e-01 -7.92745531e-01 -1.63138539e-01
-3.98965240e-01 9.12295818e-01 5.14425218e-01 3.53767723e-01
-6.39139712e-01 3.55468929e-01 6.92403495e-01 5.52591264e-01
5.86810470e-01 -6.06140256e-01 -6.78120732e-01 -1.20675385e+00
2.76569635e-01 -3.88130546e-01 -2.84302294e-01 -1.21443152e+00
-1.05992126e+00 5.13318777e-01 3.49725150e-02 -1.06564701e+00
-4.39060852e-02 -4.91334051e-01 -1.04306832e-01 9.15823042e-01
-1.34525406e+00 -1.47361124e+00 -2.69906044e-01 4.00709242e-01
4.86034870e-01 8.24689418e-02 8.87665808e-01 1.92244872e-01
1.29348516e-01 7.78508931e-02 -2.24567011e-01 -1.28964946e-01
2.97761410e-01 -9.83544827e-01 7.31398463e-01 5.26367664e-01
3.94858181e-01 2.78477013e-01 4.14455026e-01 -7.24194586e-01
-1.76915050e+00 -1.09162664e+00 5.95064521e-01 -8.53227913e-01
-4.34212014e-03 -6.83313251e-01 -1.01709330e+00 1.07685947e+00
-1.62589937e-01 3.83945048e-01 -4.96231019e-02 -3.89242083e-01
-5.62375367e-01 1.27915964e-01 -1.27670836e+00 4.62477893e-01
1.35653436e+00 -3.74367833e-01 -6.36640489e-01 2.34078825e-01
1.04334307e+00 -8.87797773e-01 -7.27484405e-01 2.67711014e-01
6.16013825e-01 -1.05969024e+00 1.53248942e+00 -4.98582035e-01
8.25254917e-01 -4.29910004e-01 -4.73144919e-01 -1.05077744e+00
-3.10435712e-01 -3.00208181e-01 -4.39064801e-01 6.95758462e-01
-1.67633500e-02 -3.36542159e-01 6.95886493e-01 5.19209862e-01
-5.13998926e-01 -9.30294991e-01 -1.00512326e+00 -5.04581511e-01
6.01386353e-02 -5.63347399e-01 7.12815344e-01 5.73710442e-01
-1.01670444e+00 2.97247171e-01 -4.63313460e-01 1.03772990e-01
8.57491910e-01 5.12626767e-01 8.90446424e-01 -1.15378428e+00
-4.37395990e-01 -3.14804882e-01 -5.07249117e-01 -1.57572353e+00
1.17883369e-01 -1.10483181e+00 5.10489242e-03 -1.77468061e+00
3.06075186e-01 -4.50086445e-01 4.83556867e-01 4.99289334e-01
4.91234601e-01 6.48317635e-01 3.20534647e-01 2.37787440e-01
-4.55332309e-01 7.87036121e-01 1.72577214e+00 -3.47122997e-02
7.31322765e-02 -2.70885468e-01 -6.12532735e-01 7.30560064e-01
1.72373369e-01 -4.18785572e-01 -2.87552595e-01 -1.10232365e+00
1.25703394e-01 3.78560394e-01 1.02596259e+00 -6.61596060e-01
-9.22948197e-02 -5.29667847e-02 7.77772307e-01 -9.07704771e-01
9.45901930e-01 -1.05062461e+00 5.52217901e-01 1.65052429e-01
-1.55998692e-01 -1.43191338e-01 2.88832188e-01 5.70173681e-01
2.01430485e-01 8.45191702e-02 9.80307996e-01 -4.49664950e-01
-3.31155956e-01 6.83917880e-01 1.97246462e-01 -3.65228131e-02
8.87643099e-01 -2.99302429e-01 -6.71681091e-02 -3.00514996e-01
-8.33972394e-01 -5.89513071e-02 9.64766800e-01 4.57673222e-01
9.77923095e-01 -1.50699377e+00 -1.00404263e+00 5.76519728e-01
-9.90690738e-02 7.72560716e-01 3.38337123e-01 4.23914343e-01
-9.83457863e-01 4.76601690e-01 -3.04159880e-01 -9.31075990e-01
-9.94096100e-01 3.08029592e-01 6.86127424e-01 -1.31643596e-04
-1.28581941e+00 8.47986460e-01 8.24888468e-01 -8.04122031e-01
2.07381323e-01 -4.79204178e-01 4.05833691e-01 -4.70982164e-01
4.12863523e-01 9.96533260e-02 2.21075639e-01 -8.83229136e-01
-1.51385486e-01 8.75408292e-01 2.03695834e-01 -2.45147049e-01
1.68553758e+00 -9.32204649e-02 -1.86641738e-01 4.90863085e-01
1.31483221e+00 -2.51325890e-02 -1.75668168e+00 -4.23309565e-01
-6.30476773e-01 -6.68232024e-01 -6.94200844e-02 -7.60947704e-01
-1.26307142e+00 8.06804359e-01 2.45181099e-01 -1.55596897e-01
6.26260877e-01 3.81489277e-01 7.06878304e-01 2.48541147e-01
5.04350245e-01 -3.02067667e-01 2.61177361e-01 6.03386819e-01
1.44455087e+00 -9.79222655e-01 2.54137427e-01 -3.18297565e-01
-2.35374704e-01 1.14040351e+00 4.76351470e-01 -5.17516017e-01
6.96250141e-01 4.37210023e-01 -4.04920667e-01 -5.11390507e-01
-7.37072766e-01 2.08419800e-01 4.27864045e-01 5.46554089e-01
2.24616081e-01 4.50173020e-02 2.98531204e-01 3.28416258e-01
-1.22275464e-01 -1.76115334e-01 4.97622490e-01 6.96642756e-01
-1.56648546e-01 -8.92205238e-01 -3.50748539e-01 4.01784956e-01
-1.64376006e-01 1.32574871e-01 -3.64275277e-01 7.41160631e-01
1.37651846e-01 1.74782857e-01 3.61066848e-01 1.17483296e-01
5.15333593e-01 -1.95566580e-01 1.12168264e+00 -8.65160942e-01
-2.00513110e-01 1.37050986e-01 -1.64346412e-01 -8.86711478e-01
-3.04898888e-01 -5.21668553e-01 -1.48289692e+00 -2.17091367e-01
1.68177187e-01 -5.40455759e-01 6.82707012e-01 9.33521390e-01
4.48862493e-01 6.53474987e-01 2.03911677e-01 -1.65107071e+00
-5.10734260e-01 -6.65784657e-01 -5.27049720e-01 4.27127838e-01
6.29524112e-01 -7.40314603e-01 -6.32901117e-02 1.78013772e-01]
|
[8.848993301391602, -3.4218969345092773]
|
77233f36-4c8a-485b-85a2-085b723a45f0
|
multi-stream-3d-fcn-with-multi-scale-deep
|
1711.10212
| null |
http://arxiv.org/abs/1711.10212v2
|
http://arxiv.org/pdf/1711.10212v2.pdf
|
Multi-stream 3D FCN with Multi-scale Deep Supervision for Multi-modality Isointense Infant Brain MR Image Segmentation
|
We present a method to address the challenging problem of segmentation of
multi-modality isointense infant brain MR images into white matter (WM), gray
matter (GM), and cerebrospinal fluid (CSF). Our method is based on
context-guided, multi-stream fully convolutional networks (FCN), which after
training, can directly map a whole volumetric data to its volume-wise labels.
In order to alleviate the poten-tial gradient vanishing problem during
training, we designed multi-scale deep supervision. Furthermore, context
infor-mation was used to further improve the performance of our method.
Validated on the test data of the MICCAI 2017 Grand Challenge on 6-month infant
brain MRI segmentation (iSeg-2017), our method achieved an average Dice Overlap
Coefficient of 95.4%, 91.6% and 89.6% for CSF, GM and WM, respectively.
|
['Guodong Zeng', 'Guoyan Zheng']
|
2017-11-28
| null | null | null | null |
['infant-brain-mri-segmentation']
|
['medical']
|
[ 2.72872239e-01 -7.93571398e-03 2.59302914e-01 -7.13594973e-01
-3.59252959e-01 -3.08847427e-01 2.10809797e-01 3.59410644e-01
-6.72750890e-01 6.54375255e-01 1.46193147e-01 -4.33698565e-01
1.62050068e-01 -4.28929061e-01 -8.74581814e-01 -4.33097363e-01
-5.43654799e-01 5.06251693e-01 2.69171655e-01 2.39466771e-01
3.63944560e-01 4.58185911e-01 -9.20524418e-01 5.49290478e-01
1.26109838e+00 9.67097998e-01 2.85157293e-01 5.57407856e-01
-2.58294582e-01 8.34396303e-01 -2.17376038e-01 -1.80294022e-01
1.26268685e-01 -1.18037976e-01 -1.10699403e+00 -1.05048850e-01
5.56811571e-01 -6.53597474e-01 -1.15171626e-01 1.21679187e+00
5.27270734e-01 2.75044646e-02 7.71023035e-01 -9.12021399e-01
-3.90981346e-01 6.65130794e-01 -8.71740818e-01 7.20790863e-01
-4.07275677e-01 -8.64257216e-02 2.15025350e-01 -8.69881749e-01
6.48114741e-01 8.75969529e-01 7.05126286e-01 6.91706359e-01
-9.53039706e-01 -8.36185336e-01 2.00266913e-01 3.14482450e-02
-9.35305834e-01 -1.29015625e-01 2.17035919e-01 -8.18820179e-01
9.22662437e-01 5.21925464e-03 6.55134141e-01 5.87944388e-01
4.27137375e-01 7.78552711e-01 1.37639499e+00 4.06373441e-02
1.39229402e-01 -5.26644647e-01 1.15096152e-01 7.61178076e-01
2.20932856e-01 -3.18078399e-01 -7.61216283e-02 -8.36370513e-02
9.03765500e-01 2.50453483e-02 -3.68879825e-01 -4.10742044e-01
-1.18138599e+00 7.65494764e-01 6.19051695e-01 3.96740496e-01
-4.66003329e-01 -1.46040380e-01 4.94663268e-01 -2.12065399e-01
9.08209503e-01 2.18300566e-01 -5.13588250e-01 1.28231987e-01
-1.24534380e+00 -1.75495148e-01 3.61142993e-01 5.42596638e-01
3.30796599e-01 -1.37662068e-01 -1.82637736e-01 9.13208306e-01
3.75164896e-01 3.57093483e-01 7.52040088e-01 -9.08944607e-01
7.76564002e-01 2.40586177e-01 -2.75582671e-01 -6.41666293e-01
-9.04927611e-01 -4.41054761e-01 -8.92048895e-01 3.74496654e-02
2.16312572e-01 -4.33648169e-01 -1.45428014e+00 1.66610241e+00
5.30418336e-01 3.34528089e-01 -3.15685093e-01 1.15337336e+00
9.79519904e-01 6.93351254e-02 1.77499458e-01 -1.33389682e-01
9.83542562e-01 -1.20586658e+00 -4.66086805e-01 -1.60490662e-01
8.47813725e-01 -4.44136441e-01 4.78614122e-01 4.82904837e-02
-1.27494407e+00 -1.68438345e-01 -9.79050398e-01 1.92417249e-01
-1.45486563e-01 -2.10830525e-01 5.93399763e-01 2.59249806e-01
-1.48988187e+00 8.30646873e-01 -1.25463617e+00 2.39523634e-01
9.10643876e-01 5.13004839e-01 -5.87916017e-01 -1.61461785e-01
-9.28349853e-01 9.68430459e-01 4.07508641e-01 -6.54242467e-03
-9.21082914e-01 -1.28769767e+00 -7.37532198e-01 -1.76760301e-01
-1.50954351e-01 -3.75656098e-01 9.35151935e-01 -6.68800890e-01
-1.14115047e+00 1.17561412e+00 6.90809041e-02 -5.90312958e-01
6.10522985e-01 -1.98654100e-01 -2.26129100e-01 8.67166638e-01
3.53479773e-01 1.04745078e+00 6.28225505e-01 -9.61084187e-01
-4.16655183e-01 -6.24447346e-01 -2.74825573e-01 9.54972953e-03
4.87798080e-02 3.75350744e-01 1.33594364e-01 -6.84147179e-01
5.70986032e-01 -7.04921782e-01 -2.98779398e-01 -1.97764248e-01
-2.87016928e-01 1.87452137e-01 4.19598162e-01 -1.41815317e+00
7.18692899e-01 -1.89424992e+00 -2.56397724e-01 2.36691624e-01
8.14201951e-01 3.14831078e-01 -2.82560915e-01 -4.87298965e-01
-5.27789652e-01 -1.47395977e-03 -7.87496328e-01 -2.97707081e-01
-4.74384785e-01 -1.13555267e-01 1.99532002e-01 7.40735054e-01
2.78409421e-01 1.00721002e+00 -1.03515673e+00 -4.34711069e-01
1.96753978e-03 6.96718454e-01 -6.14819407e-01 2.44262323e-01
2.46759504e-01 9.86369014e-01 1.01964893e-02 3.96972030e-01
9.82074916e-01 -3.50976050e-01 2.93865148e-02 9.91888810e-03
-5.32977767e-02 1.69626594e-01 -5.28489470e-01 1.85249662e+00
-3.66300702e-01 4.83890891e-01 4.76167172e-01 -1.22280002e+00
6.30599320e-01 3.05848509e-01 1.03530490e+00 -9.30312872e-01
3.01354676e-01 3.45405370e-01 4.23513979e-01 -4.46794719e-01
-4.08970341e-02 -3.46756279e-02 8.38789999e-01 6.69355989e-01
1.83367983e-01 -1.54538844e-02 2.17791751e-01 2.19713092e-01
9.85041976e-01 -4.15520146e-02 -2.36648574e-01 -7.36266255e-01
2.96589047e-01 -4.20205414e-01 6.04304433e-01 3.77051890e-01
-5.18135011e-01 1.12021184e+00 3.04677069e-01 -3.54510128e-01
-1.06834745e+00 -1.11963129e+00 -4.45403904e-01 7.69146740e-01
-2.95471877e-01 1.33823246e-01 -1.36372399e+00 -8.56338739e-01
-2.57483035e-01 3.78860176e-01 -8.11917722e-01 8.74941945e-02
-9.52850759e-01 -1.12165773e+00 1.69443041e-01 6.39344335e-01
5.32721221e-01 -9.41906989e-01 -7.62656808e-01 3.97602171e-01
-2.60268956e-01 -1.30480170e+00 -7.43426085e-01 1.85235888e-01
-1.27508378e+00 -1.14041567e+00 -1.34056365e+00 -7.78481066e-01
1.05369258e+00 1.22530656e-02 1.07271123e+00 1.50178701e-01
-3.11359733e-01 1.09102018e-01 -2.08897412e-01 -1.25205323e-01
-3.06179076e-01 -4.39942032e-02 -6.97967410e-02 -1.82568312e-01
-5.35850972e-02 -7.82069147e-01 -1.16210747e+00 1.44400477e-01
-7.93842852e-01 3.12772781e-01 2.69133508e-01 7.49261439e-01
6.25357091e-01 -5.88792861e-01 6.51155531e-01 -6.79168820e-01
3.72516394e-01 -7.86978483e-01 -4.39442813e-01 3.05432856e-01
-7.19911277e-01 -2.27741301e-01 3.24849963e-01 -5.27233064e-01
-7.35640407e-01 -1.85848191e-01 -2.48873264e-01 -3.81928563e-01
-1.21844038e-01 1.75489157e-01 4.90201175e-01 -2.60405719e-01
2.12804779e-01 1.09477833e-01 9.82417688e-02 -5.77110887e-01
1.67894125e-01 4.95213479e-01 8.00717115e-01 -5.26093781e-01
2.31618881e-02 4.64643627e-01 5.23534454e-02 -1.55508116e-01
-6.23979032e-01 -2.25761592e-01 -1.02203667e+00 -2.70758085e-02
1.38019359e+00 -7.19665527e-01 -2.45868504e-01 4.88412052e-01
-1.17275369e+00 -6.65931225e-01 1.16210878e-01 7.21536398e-01
-3.74368280e-01 3.21571648e-01 -1.07444644e+00 -2.64636338e-01
-1.00651538e+00 -1.50902009e+00 5.54187298e-01 2.06633046e-01
6.13728836e-02 -1.02128410e+00 -4.74601798e-02 5.87423384e-01
7.91385114e-01 5.19366026e-01 1.13426507e+00 -1.00447905e+00
-4.59624715e-02 1.84423864e-01 -6.99262559e-01 5.33589959e-01
1.31185010e-01 -3.27976048e-01 -7.73790836e-01 -4.94888604e-01
1.51274815e-01 -1.58090577e-01 8.51389527e-01 7.94528246e-01
1.42771220e+00 2.88771186e-02 -6.49697408e-02 9.51031625e-01
1.03131068e+00 4.07693714e-01 2.20250010e-01 1.93412170e-01
8.61924112e-01 6.34997129e-01 2.45006561e-01 2.70128667e-01
7.41099119e-01 6.91336393e-02 5.08090854e-01 -3.45373839e-01
-5.15978158e-01 1.03302866e-01 -2.01496229e-01 1.20229423e+00
2.73402557e-02 4.32617545e-01 -1.26708317e+00 6.94329262e-01
-1.33790755e+00 -2.50927329e-01 -9.04638320e-02 1.86162448e+00
1.05671167e+00 7.00937875e-04 -3.96439619e-02 -3.03449363e-01
8.74162614e-01 -9.94675383e-02 -7.34017015e-01 -4.31786984e-01
-1.11970520e-02 5.69724858e-01 5.63361764e-01 3.32506627e-01
-1.24955559e+00 6.74723566e-01 6.43172026e+00 3.82448733e-01
-1.49753165e+00 5.88347793e-01 1.19599652e+00 -1.19685389e-01
-5.69599643e-02 -6.03568017e-01 -2.55217403e-01 6.45122111e-01
1.02947807e+00 1.71409205e-01 5.33359051e-01 5.47440946e-01
1.64353829e-02 -2.14786351e-01 -8.47960472e-01 7.28410363e-01
-4.40437645e-02 -1.24382889e+00 -4.20022398e-01 -3.32128137e-01
1.11101890e+00 8.07432771e-01 6.74923137e-02 5.02544045e-02
2.89001204e-02 -1.31379735e+00 5.40112555e-01 3.64934087e-01
1.13880634e+00 -7.15274751e-01 7.21087456e-01 1.82255253e-01
-8.36378157e-01 2.08751991e-01 -1.07106559e-01 2.92375147e-01
4.70139980e-02 7.23632753e-01 -8.97397637e-01 2.31042057e-01
8.24885726e-01 4.16201681e-01 -4.17334765e-01 1.24362361e+00
-1.05408773e-01 5.32387376e-01 -2.66670763e-01 4.38072175e-01
3.85851175e-01 -1.95148572e-01 2.09485963e-01 1.31632257e+00
8.00822750e-02 3.31124604e-01 9.10003185e-02 9.61131871e-01
-2.84622312e-01 3.37463021e-01 -4.75809649e-02 2.27825567e-01
1.86976492e-02 1.24793434e+00 -1.11946750e+00 -5.28917134e-01
-4.03419733e-01 7.28412092e-01 4.47276711e-01 3.15142274e-01
-6.07429206e-01 -1.37406334e-01 3.38729382e-01 6.40197098e-02
2.51412570e-01 -1.68534085e-01 -7.43141115e-01 -1.24231076e+00
4.65634791e-03 -6.38769448e-01 1.03104413e-01 -4.48057294e-01
-1.05211341e+00 8.55549693e-01 -4.07590438e-03 -5.43996990e-01
1.33847296e-02 -4.41824198e-01 -7.10228503e-01 1.09087265e+00
-1.76128852e+00 -6.60421133e-01 -3.62158939e-02 3.61969650e-01
1.62493959e-01 -1.26157582e-01 5.36785245e-01 6.05732918e-01
-4.63102609e-01 5.89166999e-01 3.65464278e-02 4.64883476e-01
3.34408969e-01 -1.22595644e+00 5.18322825e-01 9.31964338e-01
-4.01615500e-01 7.97233522e-01 2.63919741e-01 -9.42706347e-01
-6.59387290e-01 -1.25460160e+00 5.11413991e-01 9.02358294e-02
7.28998542e-01 -1.59651741e-01 -1.14600289e+00 5.18354714e-01
7.27602392e-02 4.84923095e-01 5.28078973e-01 -4.01874393e-01
-2.68191814e-01 4.19772387e-01 -1.67921162e+00 2.64004290e-01
9.97414649e-01 -2.64569581e-01 -6.08513594e-01 5.13833225e-01
9.90339339e-01 -8.99249554e-01 -1.25431395e+00 9.02801692e-01
2.34387785e-01 -8.75161946e-01 9.28441644e-01 -6.78747535e-01
7.98468947e-01 1.31270662e-01 8.01830366e-02 -1.37246656e+00
-1.59801561e-02 -1.63472027e-01 -2.63838977e-01 6.22274041e-01
3.14815342e-01 -7.98789799e-01 5.98369956e-01 7.72666097e-01
-4.71376479e-01 -1.29303348e+00 -9.61772025e-01 -3.47620279e-01
7.71955907e-01 -2.50237912e-01 8.22088420e-01 1.04873335e+00
-3.62738371e-01 -3.18802208e-01 1.15712419e-01 4.34414670e-02
8.69501948e-01 -2.07988583e-02 -2.78604180e-01 -9.20211792e-01
1.77459717e-01 -5.96529901e-01 -2.12051243e-01 -5.81512868e-01
2.07014874e-01 -1.30078626e+00 2.17132866e-02 -1.63040698e+00
5.33129275e-01 -4.23248827e-01 -7.99689651e-01 5.62662482e-01
-3.26726496e-01 3.13358128e-01 1.46856040e-01 -1.06934510e-01
-5.22761106e-01 1.77688494e-01 1.65254891e+00 -6.51901811e-02
1.21282794e-01 -1.82046905e-01 -3.22613001e-01 7.43592143e-01
9.99150753e-01 -4.12941843e-01 -3.26989412e-01 -8.17220271e-01
-4.46602941e-01 1.66664749e-01 6.67684004e-02 -7.68883944e-01
-5.57842255e-02 5.82018010e-02 6.10915959e-01 -6.43516004e-01
-1.02875866e-01 -3.63207072e-01 -4.85519975e-01 6.25118375e-01
-3.45525146e-01 2.57620454e-01 1.59593567e-01 -2.48230755e-01
-1.01737238e-01 -5.82616888e-02 1.12761390e+00 -1.64816126e-01
-2.55594105e-01 1.00850391e+00 -1.70948654e-01 5.10052860e-01
6.61968887e-01 2.44685501e-01 -2.14730740e-01 7.84472004e-02
-8.91673028e-01 4.45087045e-01 2.31856450e-01 3.34709913e-01
8.12756896e-01 -9.93416369e-01 -7.80936360e-01 2.46133611e-01
-3.36127788e-01 1.24629952e-01 4.09217149e-01 1.37951982e+00
-8.30110013e-01 4.35367346e-01 -5.08400142e-01 -7.49642134e-01
-9.31847811e-01 1.58970505e-01 5.58888137e-01 -2.62829989e-01
-9.09551620e-01 1.21134603e+00 2.99609780e-01 -6.15378737e-01
2.34980002e-01 -6.33943677e-01 -2.78984755e-01 -3.17656070e-01
8.44083726e-01 3.34583014e-01 2.98518866e-01 -7.24858403e-01
-4.67466474e-01 4.73759711e-01 -2.76733875e-01 -3.59466448e-02
1.65683770e+00 -1.77208006e-01 -5.82625091e-01 7.88411722e-02
1.51964724e+00 -6.11904800e-01 -1.35205400e+00 -1.89117715e-01
2.47165620e-01 -2.59803236e-01 2.74140656e-01 -1.03502631e+00
-1.58075619e+00 1.27753496e+00 1.04220295e+00 -2.66849875e-01
8.78947139e-01 -3.12481344e-01 1.18679202e+00 -3.29361737e-01
4.15417641e-01 -9.72140729e-01 -2.14991361e-01 5.81360459e-01
8.30756783e-01 -1.50398457e+00 -2.90310472e-01 -5.73451966e-02
-4.91413534e-01 1.20267403e+00 6.86911523e-01 -1.97077036e-01
7.35826433e-01 4.71960008e-01 2.62611657e-01 -2.50084817e-01
-4.60158169e-01 1.71231821e-01 5.43097615e-01 5.90942979e-01
6.14963651e-01 2.75265008e-01 -4.78083611e-01 5.60695946e-01
-9.96175334e-02 -4.05385047e-02 2.58209318e-01 7.65760064e-01
-4.16342169e-01 -5.74417233e-01 -1.60485417e-01 8.87038708e-01
-9.85881150e-01 -4.73133832e-01 3.22127283e-01 2.64770418e-01
2.80583858e-01 7.69075215e-01 1.09951623e-01 -6.71286583e-02
-2.57300943e-01 -4.06101532e-03 6.01589441e-01 -6.47585392e-01
-6.01651013e-01 -4.50253412e-02 -2.00068921e-01 -7.59595156e-01
-1.84207857e-01 -6.33712232e-01 -1.92591345e+00 -9.26132202e-02
1.47000805e-01 -3.38648893e-02 1.27216148e+00 1.28283703e+00
1.87589616e-01 6.66646540e-01 5.38887739e-01 -7.09292412e-01
-1.66365653e-01 -1.06278038e+00 -4.49940920e-01 4.80333954e-01
5.50195277e-01 -5.40230513e-01 -1.30688652e-01 -2.94142574e-01]
|
[14.188393592834473, -2.363065004348755]
|
188a360e-30a2-4910-95ba-c5b0c04eee06
|
ldmic-learning-based-distributed-multi-view
|
2301.09799
| null |
https://arxiv.org/abs/2301.09799v3
|
https://arxiv.org/pdf/2301.09799v3.pdf
|
LDMIC: Learning-based Distributed Multi-view Image Coding
|
Multi-view image compression plays a critical role in 3D-related applications. Existing methods adopt a predictive coding architecture, which requires joint encoding to compress the corresponding disparity as well as residual information. This demands collaboration among cameras and enforces the epipolar geometric constraint between different views, which makes it challenging to deploy these methods in distributed camera systems with randomly overlapping fields of view. Meanwhile, distributed source coding theory indicates that efficient data compression of correlated sources can be achieved by independent encoding and joint decoding, which motivates us to design a learning-based distributed multi-view image coding (LDMIC) framework. With independent encoders, LDMIC introduces a simple yet effective joint context transfer module based on the cross-attention mechanism at the decoder to effectively capture the global inter-view correlations, which is insensitive to the geometric relationships between images. Experimental results show that LDMIC significantly outperforms both traditional and learning-based MIC methods while enjoying fast encoding speed. Code will be released at https://github.com/Xinjie-Q/LDMIC.
|
['Jun Zhang', 'Jiawei Shao', 'Xinjie Zhang']
|
2023-01-24
| null | null | null | null |
['data-compression']
|
['time-series']
|
[-7.43743926e-02 -3.27054977e-01 -2.42289722e-01 -2.62528211e-01
-6.62460446e-01 -3.12724352e-01 3.05891007e-01 -1.73884168e-01
1.12712048e-01 2.57391661e-01 4.27558333e-01 -2.01350097e-02
-1.05991788e-01 -7.09496200e-01 -7.80854404e-01 -7.00590193e-01
7.67703429e-02 -1.28462417e-02 1.88473001e-01 3.04183085e-02
3.52141052e-01 -3.22128907e-02 -1.40119767e+00 3.80853027e-01
7.82927752e-01 1.06765676e+00 8.33949745e-01 6.33019567e-01
1.27134040e-01 1.20736468e+00 -2.16642737e-01 -3.61457765e-01
3.62599373e-01 -5.60020447e-01 -3.87971610e-01 2.48163447e-01
3.61976206e-01 -8.58576477e-01 -7.69737542e-01 1.11964428e+00
5.43865621e-01 -1.88578784e-01 1.95305184e-01 -1.08408296e+00
-8.79107296e-01 3.16022515e-01 -8.68133008e-01 1.51045665e-01
4.35300440e-01 -8.83976743e-02 9.76507545e-01 -7.73410499e-01
3.46931726e-01 1.06398511e+00 3.04534853e-01 1.84108645e-01
-8.03954542e-01 -6.11518443e-01 2.18674600e-01 4.65478718e-01
-1.59850395e+00 -5.82593083e-01 9.68554616e-01 -6.98021576e-02
6.75583661e-01 5.06600700e-02 6.52446985e-01 5.83943665e-01
2.94377208e-01 8.46230805e-01 8.66083622e-01 -2.30359674e-01
1.94455147e-01 -2.02895343e-01 -5.49000502e-01 6.32052541e-01
2.27293655e-01 1.25267863e-01 -9.13816810e-01 4.29492965e-02
1.04427385e+00 4.04006749e-01 -5.71358144e-01 -5.50557435e-01
-1.25940537e+00 7.37128854e-01 4.47073489e-01 1.03820145e-01
-2.54562259e-01 2.76864886e-01 2.80451655e-01 2.32137263e-01
5.29052019e-01 -3.37621927e-01 -2.99385577e-01 -1.28189147e-01
-7.79199958e-01 -2.39251167e-01 5.15650332e-01 1.38961112e+00
7.94346213e-01 -2.32282847e-01 2.79226184e-01 1.02834427e+00
4.73473161e-01 6.24105573e-01 2.68200517e-01 -1.34680665e+00
7.63272047e-01 5.67940235e-01 -2.39759237e-01 -1.43734574e+00
1.71054199e-01 -2.36862466e-01 -1.21761084e+00 -1.31897762e-01
-3.00228328e-01 2.59506982e-02 -3.38006973e-01 1.56044972e+00
2.65109420e-01 4.19475764e-01 -6.93856627e-02 1.06612539e+00
5.35562873e-01 8.14115584e-01 -6.25498950e-01 -4.42388266e-01
8.99268568e-01 -1.11729288e+00 -5.37790298e-01 -1.72882468e-01
3.56007606e-01 -9.16243672e-01 6.23071015e-01 3.98227543e-01
-1.34850287e+00 -5.33022642e-01 -1.08009839e+00 -3.30189168e-01
2.55913287e-01 3.70493494e-02 5.47403753e-01 3.08890671e-01
-1.10351419e+00 3.49133536e-02 -9.15846288e-01 5.72901443e-02
5.71178138e-01 3.33391696e-01 -2.90288985e-01 -9.42948937e-01
-7.42516100e-01 4.06170875e-01 -1.81205366e-02 -1.11679852e-01
-7.76485026e-01 -4.99404550e-01 -8.72079909e-01 1.00823060e-01
2.35059991e-01 -8.66661906e-01 1.01350009e+00 -7.55388021e-01
-1.33609438e+00 5.48328698e-01 -4.82430816e-01 -5.68108521e-02
8.93720612e-02 -3.07781667e-01 -2.01320112e-01 5.65025866e-01
2.06349954e-01 5.25712132e-01 7.25825787e-01 -1.35365903e+00
-6.18083119e-01 -3.32923740e-01 2.04576075e-01 5.75058818e-01
-4.33146924e-01 -1.41296133e-01 -1.11296213e+00 -5.13622820e-01
3.95546705e-01 -8.33450317e-01 -1.54411450e-01 2.41931081e-01
-2.03658462e-01 2.48822570e-01 7.93795466e-01 -4.82482702e-01
1.26953661e+00 -2.31103277e+00 3.59333724e-01 -7.69667476e-02
5.11176765e-01 -4.24530767e-02 -1.97360247e-01 5.64624012e-01
1.64947510e-01 -3.94909054e-01 -9.73727405e-02 -4.71380800e-01
-3.41119915e-01 2.59525120e-01 -1.23806886e-01 6.38003469e-01
-2.77831852e-01 5.86126864e-01 -8.67831409e-01 -4.99360263e-01
4.77863461e-01 6.92847013e-01 -9.73474085e-01 4.19035524e-01
9.36787724e-02 5.61622977e-01 -5.26965439e-01 6.13871694e-01
1.15916038e+00 -7.08737969e-01 3.27279210e-01 -2.85462707e-01
-1.76385064e-02 1.42115027e-01 -1.01841450e+00 2.18537354e+00
-7.21272051e-01 5.38137138e-01 1.25384748e-01 -1.10872376e+00
7.42171228e-01 2.88086325e-01 6.85882568e-01 -6.34671688e-01
-1.11428788e-03 2.08669633e-01 -2.86893159e-01 -3.32386941e-01
3.40682924e-01 1.38794169e-01 1.03512727e-01 4.45924759e-01
-7.14362189e-02 1.12554347e-02 -1.29473299e-01 5.49624801e-01
1.02753925e+00 -5.75051121e-02 2.61464924e-01 1.16985708e-01
5.54567575e-01 -4.77701515e-01 8.18019569e-01 2.35891595e-01
-3.65274064e-02 9.41324174e-01 1.60314530e-01 -4.14567322e-01
-1.02213645e+00 -9.51124132e-01 2.01579452e-01 5.67903042e-01
8.37052882e-01 -7.40006387e-01 -3.61604422e-01 -1.89820647e-01
-3.22466463e-01 2.82141417e-01 -1.38679028e-01 -2.66919553e-01
-4.81273234e-01 -3.89994085e-01 -1.53675064e-01 3.56970549e-01
7.83451796e-01 -1.98056355e-01 -6.02833331e-01 1.33569241e-01
-6.30355537e-01 -1.35834956e+00 -7.66141176e-01 -8.50435197e-02
-9.21845496e-01 -1.10283959e+00 -8.44082296e-01 -7.37812757e-01
7.30708003e-01 1.19314373e+00 1.07645595e+00 2.82242328e-01
4.85306233e-02 6.02252901e-01 -6.31689966e-01 3.24280523e-02
-1.54580563e-01 -1.84658363e-01 -1.80285662e-01 7.57488310e-02
3.48024189e-01 -1.02627742e+00 -1.02151859e+00 4.27167743e-01
-8.93988848e-01 5.03877938e-01 6.82167470e-01 8.56506884e-01
7.41229117e-01 1.37226656e-01 1.40675858e-01 -3.68090212e-01
-5.86102949e-03 -6.49424791e-01 -6.10273719e-01 2.39962175e-01
-3.80578578e-01 -2.81809479e-01 7.05166399e-01 3.17749083e-02
-9.33127642e-01 8.27728584e-02 4.79213707e-02 -8.47692430e-01
5.92373461e-02 3.74901593e-01 -4.48348790e-01 -1.95243523e-01
-7.35318363e-02 6.55431390e-01 -1.23929139e-02 -2.89061308e-01
2.80271679e-01 6.99538469e-01 2.78345674e-01 -2.33163908e-01
5.69366455e-01 7.02727258e-01 -1.13245599e-01 -7.07966685e-01
-7.13087738e-01 -4.83916700e-01 -5.28755605e-01 -2.71476597e-01
6.20162845e-01 -1.70704484e+00 -4.71184790e-01 4.49292570e-01
-1.31142092e+00 -1.00593179e-01 1.65670887e-01 7.23141730e-01
-7.44705856e-01 6.03264272e-01 -7.27075934e-01 -4.15068299e-01
-4.60062660e-02 -1.27824938e+00 1.16383886e+00 1.30419225e-01
3.59786630e-01 -9.04805899e-01 -2.27016807e-02 5.97962081e-01
3.47501040e-01 -2.84478992e-01 5.71402967e-01 1.70317039e-01
-1.33564782e+00 -6.60046563e-02 -3.81368220e-01 3.74294996e-01
3.45931083e-01 -3.51848394e-01 -7.29612291e-01 -4.36248422e-01
2.20960826e-01 -2.42514402e-01 7.56552994e-01 5.81552565e-01
1.32216024e+00 -1.89694270e-01 -3.38629931e-01 1.08326137e+00
1.76479483e+00 1.98617190e-01 7.63627946e-01 -2.01167930e-02
9.31299686e-01 2.04443380e-01 3.69526893e-01 8.82992566e-01
9.49783325e-01 6.38227403e-01 6.28007591e-01 1.25406817e-01
-2.72057146e-01 -4.31194991e-01 4.34549868e-01 1.50223649e+00
-1.52894007e-02 -5.53608656e-01 -5.87897480e-01 5.46672523e-01
-1.83810210e+00 -1.01494312e+00 3.16853710e-02 2.27190351e+00
5.39729953e-01 -3.22192222e-01 -4.39290851e-01 -5.90379685e-02
8.25759053e-01 4.84334618e-01 -5.56548476e-01 -4.71786000e-02
-2.53172487e-01 -2.16773704e-01 4.99455839e-01 3.96689445e-01
-9.50904012e-01 4.89890814e-01 5.22867966e+00 9.46434319e-01
-9.28640068e-01 3.13037753e-01 8.80859613e-01 -4.07347828e-01
-4.97011960e-01 6.47910386e-02 -5.50535262e-01 7.51613498e-01
7.30511427e-01 -1.42966017e-01 5.73010385e-01 7.09155619e-01
1.94982052e-01 -3.76075745e-01 -8.22230399e-01 1.50090814e+00
3.68040323e-01 -1.61257613e+00 1.69718415e-02 3.87252241e-01
1.10260880e+00 3.45635027e-01 -6.67545125e-02 -3.74105573e-01
1.52090386e-01 -5.35622299e-01 5.36363661e-01 2.40308508e-01
9.63863671e-01 -8.84524286e-01 5.24244666e-01 4.33166087e-01
-1.48461521e+00 -2.33786762e-01 -6.85879290e-01 4.07163911e-02
3.42934668e-01 7.88547039e-01 7.72941560e-02 9.13829684e-01
8.07610273e-01 1.31817150e+00 -2.28394121e-01 9.16429400e-01
-1.23508506e-01 2.37637296e-01 -3.45534712e-01 4.72272813e-01
-7.51735121e-02 -3.18644524e-01 4.03261781e-01 6.04221761e-01
8.09462130e-01 3.20480317e-01 1.50257885e-01 4.55048174e-01
-1.11333795e-01 -1.10089527e-02 -8.26863945e-01 3.25655758e-01
6.28842115e-01 9.29232359e-01 -4.60648865e-01 -2.21829921e-01
-1.08750629e+00 1.39497805e+00 2.13849157e-01 1.65288791e-01
-7.36551702e-01 -1.70644429e-02 6.01068735e-01 -2.58235354e-03
7.24635422e-01 -4.79957074e-01 -2.21416011e-01 -1.48801053e+00
2.19502792e-01 -8.19005311e-01 2.33005062e-01 -8.48495543e-01
-1.09222329e+00 4.35310781e-01 -1.85970485e-01 -1.92408371e+00
-1.73573848e-02 -1.34625614e-01 -4.58807260e-01 6.10633552e-01
-1.85381448e+00 -1.16427755e+00 -5.11274397e-01 8.82385910e-01
5.85410237e-01 -1.68811888e-01 6.07961595e-01 5.83503723e-01
-3.82522851e-01 4.45122719e-01 2.84574777e-01 -2.91038677e-02
7.41897047e-01 -7.48424411e-01 -3.93270999e-02 9.34727371e-01
9.32825729e-02 3.70781571e-01 4.50858288e-02 -3.50135088e-01
-1.90340781e+00 -1.04448152e+00 7.92125165e-01 1.07870616e-01
2.13816613e-01 -1.10764183e-01 -8.00576389e-01 4.91468012e-01
5.61329782e-01 3.94017190e-01 9.82784808e-01 -3.30775529e-01
-3.98012400e-01 -2.88343579e-01 -6.39622092e-01 2.80105770e-01
1.20082486e+00 -6.68226480e-01 -2.52277367e-02 4.07530844e-01
7.37499177e-01 -3.50427419e-01 -7.64070928e-01 1.09208211e-01
3.32667679e-01 -1.72293651e+00 1.00849009e+00 4.08371419e-01
1.09897447e+00 -4.20652181e-01 -7.38762021e-01 -1.07389009e+00
-3.31457704e-01 -5.62180042e-01 -4.42216605e-01 1.10552990e+00
-1.24634586e-01 -5.24346769e-01 6.22806787e-01 3.46836597e-01
-1.82950720e-01 -8.31738651e-01 -9.38425839e-01 -6.76985323e-01
-2.15631098e-01 -3.82748961e-01 5.19112647e-01 8.56055856e-01
-4.29959036e-03 2.82988518e-01 -7.06712961e-01 5.58239341e-01
8.64143372e-01 5.34435928e-01 7.82827020e-01 -6.55681610e-01
-5.98076105e-01 -1.02013357e-01 -5.10874033e-01 -1.85390389e+00
-1.19758032e-01 -7.66569197e-01 -1.08497411e-01 -1.42955148e+00
5.41629791e-01 -5.41399181e-01 -1.59410611e-01 1.09438384e-02
-6.84514940e-02 2.71292895e-01 2.47744769e-01 6.85974956e-01
-9.28628445e-01 1.00133300e+00 1.31241131e+00 1.16700366e-01
1.44015878e-01 -2.20890298e-01 -8.65684986e-01 6.54008389e-01
8.07851851e-01 -2.88583547e-01 -7.73851275e-01 -1.16208816e+00
2.41211981e-01 4.98138785e-01 2.94254035e-01 -1.06002176e+00
7.03691185e-01 -1.39307722e-01 3.40476453e-01 -6.72334552e-01
4.88596380e-01 -9.85491157e-01 1.98041603e-01 2.86919922e-01
-1.90585498e-02 1.94894224e-01 -2.48128191e-01 8.12866271e-01
-7.08721101e-01 1.80080056e-01 6.58843756e-01 -2.25557551e-01
-7.94888794e-01 6.90825164e-01 -8.73270631e-02 -4.60211374e-02
9.75202441e-01 -2.50993699e-01 -2.63510823e-01 -7.75434971e-01
-3.53782624e-02 2.65407890e-01 8.04539561e-01 2.80690253e-01
1.04340982e+00 -1.37896633e+00 -6.84230268e-01 5.34300745e-01
1.54294401e-01 3.11429977e-01 7.81381607e-01 9.18489099e-01
-5.92436671e-01 2.90126354e-01 -1.83716081e-02 -7.76479185e-01
-1.16492224e+00 5.91175377e-01 -8.73728618e-02 2.46179085e-02
-6.70544147e-01 9.05877292e-01 5.42961121e-01 -3.17735560e-02
2.42924225e-02 6.17204979e-02 1.87389195e-01 -3.91201288e-01
6.82441473e-01 1.49057522e-01 -2.52604634e-01 -6.16087735e-01
-2.53655434e-01 8.57011855e-01 -2.09023617e-02 2.06028119e-01
1.44124222e+00 -8.42177510e-01 -1.93989694e-01 2.02765882e-01
1.61433184e+00 7.23711476e-02 -1.45392156e+00 -4.13480729e-01
-6.59672499e-01 -1.07525229e+00 3.74648094e-01 -2.54050851e-01
-1.55264282e+00 9.21348214e-01 4.03364331e-01 -1.60582989e-01
1.61119258e+00 7.78231621e-02 1.19686604e+00 -1.27599854e-02
8.06203783e-01 -8.14864993e-01 2.65345514e-01 3.37544620e-01
7.67136872e-01 -1.38963878e+00 2.87145108e-01 -6.98212504e-01
-5.62267005e-01 9.86912608e-01 5.53227663e-01 -8.75760689e-02
9.08366024e-01 3.15469503e-01 -1.78615987e-01 -9.96038243e-02
-1.16798306e+00 1.08571492e-01 -3.84437814e-02 6.09544992e-01
3.55112433e-01 -1.48453221e-01 2.70037968e-02 2.12980822e-01
2.81114370e-01 8.51953477e-02 4.91091013e-01 9.20413494e-01
-2.07461357e-01 -1.28698945e+00 -1.96585625e-01 2.43074343e-01
-1.97334900e-01 -2.65166342e-01 1.57952800e-01 3.10356081e-01
2.41015628e-01 1.02380097e+00 2.33160794e-01 -6.39768958e-01
-1.84475332e-01 -6.23703778e-01 4.98510897e-01 -5.35786510e-01
1.65593222e-01 4.78283554e-01 -3.32886249e-01 -8.45550239e-01
-9.31718886e-01 -7.25860059e-01 -8.84937823e-01 -5.43052137e-01
-2.90680677e-01 -9.27428231e-02 3.17516357e-01 5.98616719e-01
8.76381278e-01 2.28824362e-01 1.27089548e+00 -8.35782707e-01
-1.15148000e-01 -3.95247847e-01 -6.05214775e-01 7.55631253e-02
3.93846273e-01 -4.56412852e-01 -3.52405012e-01 3.86004061e-01]
|
[11.01176643371582, -1.7584235668182373]
|
52ce978e-4ed6-45f6-a71d-d101476e0b6a
|
a-permutation-free-kernel-two-sample-test
|
2211.14908
| null |
https://arxiv.org/abs/2211.14908v2
|
https://arxiv.org/pdf/2211.14908v2.pdf
|
A Permutation-free Kernel Two-Sample Test
|
The kernel Maximum Mean Discrepancy~(MMD) is a popular multivariate distance metric between distributions that has found utility in two-sample testing. The usual kernel-MMD test statistic is a degenerate U-statistic under the null, and thus it has an intractable limiting distribution. Hence, to design a level-$\alpha$ test, one usually selects the rejection threshold as the $(1-\alpha)$-quantile of the permutation distribution. The resulting nonparametric test has finite-sample validity but suffers from large computational cost, since every permutation takes quadratic time. We propose the cross-MMD, a new quadratic-time MMD test statistic based on sample-splitting and studentization. We prove that under mild assumptions, the cross-MMD has a limiting standard Gaussian distribution under the null. Importantly, we also show that the resulting test is consistent against any fixed alternative, and when using the Gaussian kernel, it has minimax rate-optimal power against local alternatives. For large sample sizes, our new cross-MMD provides a significant speedup over the MMD, for only a slight loss in power.
|
['Aaditya Ramdas', 'Ilmun Kim', 'Shubhanshu Shekhar']
|
2022-11-27
| null | null | null | null |
['hypothesis-testing', 'hypothesis-testing']
|
['methodology', 'miscellaneous']
|
[ 7.03617185e-02 -1.09449692e-01 -4.78810966e-01 -3.66885543e-01
-1.08393836e+00 -6.99736357e-01 6.91204369e-02 2.44487941e-01
-4.64241832e-01 1.15327573e+00 -5.48184216e-01 -8.41822743e-01
-5.11822462e-01 -9.30757642e-01 -7.32776284e-01 -8.28993082e-01
-3.89700174e-01 5.20348728e-01 3.03646743e-01 4.78580266e-01
5.10801256e-01 3.31911266e-01 -1.44025862e+00 -4.66410309e-01
1.47717392e+00 9.32837784e-01 -2.19988286e-01 7.67560124e-01
3.48682314e-01 -8.15394893e-02 -5.94119251e-01 -2.76693881e-01
3.66078317e-01 -7.36323714e-01 -5.59173584e-01 -4.71822858e-01
6.17309630e-01 -2.05044970e-01 1.81794047e-01 1.42734337e+00
7.78498054e-01 1.85098290e-01 9.61649776e-01 -1.57992649e+00
-5.42585552e-01 4.09086764e-01 -9.87807870e-01 7.49310851e-02
3.67498279e-01 -7.67699555e-02 8.97918403e-01 -9.07941222e-01
1.84739575e-01 1.15589893e+00 8.75519872e-01 1.39934808e-01
-1.88853073e+00 -7.08340824e-01 -5.97568750e-01 -2.19750136e-01
-1.78779066e+00 -1.97781399e-02 1.60229370e-01 -3.80678117e-01
4.20718491e-01 5.61577559e-01 3.34211379e-01 6.57756090e-01
6.43892109e-01 3.77668411e-01 1.60848868e+00 -3.90515357e-01
6.30883276e-01 5.41314445e-02 7.00682327e-02 6.53059244e-01
7.92887628e-01 -1.41606301e-01 -2.33853653e-01 -8.18599343e-01
9.45711434e-01 -3.52011472e-01 -2.56379336e-01 -5.72110713e-01
-1.03642452e+00 8.68946671e-01 -5.10949314e-01 1.55390248e-01
9.69116986e-02 5.97395189e-02 9.76229087e-02 4.71601725e-01
5.31927526e-01 2.06779167e-01 -5.32806337e-01 -2.28720918e-01
-1.02671182e+00 4.82244998e-01 9.92195308e-01 7.47024596e-01
5.45827329e-01 -1.38795882e-01 -4.08351064e-01 6.18297219e-01
-3.12965401e-02 1.19880009e+00 5.44317625e-02 -8.37469697e-01
2.30824649e-01 5.38462512e-02 4.42025453e-01 -8.25529635e-01
-3.72683883e-01 -4.82738584e-01 -9.63627756e-01 3.08019757e-01
7.08110750e-01 -1.72971517e-01 -4.50017959e-01 2.01937842e+00
3.33736360e-01 3.54557395e-01 -3.64873081e-01 4.35288697e-01
-9.59639326e-02 3.84827137e-01 -2.10726425e-01 -6.26671612e-01
1.01587141e+00 -4.52415757e-02 -4.89976645e-01 1.18013188e-01
9.20963168e-01 -5.85598469e-01 1.12634933e+00 4.93746310e-01
-1.19868529e+00 -2.50084251e-01 -1.06448257e+00 3.58430266e-01
-3.76790017e-02 -2.00782686e-01 3.61438960e-01 9.51228857e-01
-1.22110188e+00 6.03821218e-01 -5.60465038e-01 -2.79140145e-01
1.30910963e-01 4.82889801e-01 -6.22784615e-01 -2.55564302e-01
-7.71060169e-01 6.07078910e-01 2.93754283e-02 -2.45059550e-01
-3.52166325e-01 -9.48339999e-01 -7.59759843e-01 6.15022667e-02
2.59157211e-01 -3.49380404e-01 8.95707965e-01 -1.83510736e-01
-1.18872392e+00 7.04751730e-01 -2.98400640e-01 -1.15396708e-01
5.35187304e-01 1.22660749e-01 -3.25462639e-01 -6.93749562e-02
6.43485546e-01 1.63766891e-01 6.19517982e-01 -7.16300130e-01
-6.60727620e-01 -5.49654782e-01 -5.56736290e-01 -1.44091338e-01
9.10754651e-02 -1.90351367e-01 -1.16321713e-01 -5.81980586e-01
2.65370935e-01 -8.09844255e-01 -6.39699847e-02 -2.61507183e-01
-4.88029391e-01 -3.78937304e-01 3.08637112e-01 -3.16593885e-01
1.46788776e+00 -2.45536160e+00 -4.91490066e-01 8.91475618e-01
1.52542852e-02 -1.60912439e-01 -7.61879310e-02 1.37542173e-01
-2.12537289e-01 1.57224685e-01 -5.10194302e-01 4.00620311e-01
2.59662658e-01 -2.08581015e-01 -1.12046979e-01 1.06563401e+00
1.36096757e-02 5.81358373e-01 -8.36994529e-01 -3.26794118e-01
-3.39670986e-01 -1.82280988e-02 -8.30673516e-01 1.09756095e-02
3.64289284e-01 -5.90200648e-02 -1.24983929e-01 5.05794704e-01
1.37449765e+00 -1.83624014e-01 2.56928295e-01 8.34473372e-02
-2.44333968e-01 -2.54142016e-01 -1.38951206e+00 9.53070641e-01
1.20177358e-01 5.57401836e-01 3.73297632e-02 -1.20993865e+00
1.01702309e+00 -9.58021507e-02 2.49509141e-01 -5.32786727e-01
9.31954086e-02 7.97083378e-01 7.02017546e-02 -2.97036648e-01
4.20821719e-02 -3.57158631e-01 -3.70810151e-01 4.20776606e-01
-1.77955702e-01 -2.64832936e-02 1.06491596e-01 -3.12905878e-01
1.37721527e+00 -2.90939331e-01 6.57205462e-01 -1.10963285e+00
2.72791594e-01 -2.94489741e-01 7.93195307e-01 1.21373320e+00
-2.08088353e-01 4.08838779e-01 1.26617837e+00 2.69702166e-01
-5.96245766e-01 -1.70777845e+00 -7.52181113e-01 7.94992328e-01
3.97715457e-02 2.78425694e-01 -6.54639840e-01 -6.31035984e-01
6.76667631e-01 8.40291142e-01 -7.62714267e-01 -1.48369253e-01
-2.80325085e-01 -8.89989913e-01 7.41106331e-01 3.04676622e-01
4.18255061e-01 -2.00454548e-01 -2.53589243e-01 -2.99175233e-02
1.42220765e-01 -5.32616258e-01 -7.36306489e-01 4.05375570e-01
-9.71541822e-01 -1.29000294e+00 -8.11946154e-01 -9.12703335e-01
7.83605635e-01 1.90504268e-01 7.71283209e-01 -5.55125892e-01
-1.79382801e-01 3.00355405e-01 4.94501628e-02 5.13894763e-03
-1.41229061e-02 -4.59220022e-01 3.42734277e-01 -1.94319025e-01
6.41665936e-01 -5.56543291e-01 -6.68207407e-01 6.06690109e-01
-8.94866943e-01 -7.54214764e-01 5.85118711e-01 9.43181336e-01
7.69101858e-01 3.34000170e-01 8.20944250e-01 -7.79251635e-01
7.52178729e-01 -5.93914568e-01 -9.68860090e-01 1.82948440e-01
-6.60226524e-01 1.85390964e-01 5.24195254e-01 -5.19651353e-01
-4.20362025e-01 -7.29591787e-01 3.41997027e-01 -9.21325684e-02
4.27293703e-02 5.64041495e-01 -3.44820648e-01 -1.21154830e-01
5.90429246e-01 1.63728669e-01 1.41573370e-01 -3.82458925e-01
-2.23237574e-02 5.52530885e-01 7.68379152e-01 -8.14167976e-01
7.64096022e-01 2.62793392e-01 6.02827251e-01 -9.11124885e-01
-6.27061486e-01 -3.62020552e-01 -3.48774523e-01 2.74596572e-01
6.77250922e-01 -4.28304255e-01 -1.17066789e+00 2.73222536e-01
-5.76827288e-01 -4.64631855e-01 -3.25322747e-01 1.00437319e+00
-7.30095267e-01 4.76015806e-01 -1.28174931e-01 -9.16010439e-01
1.13031954e-01 -8.85979354e-01 7.18472004e-01 5.04230298e-02
-1.94273055e-01 -9.03580606e-01 2.39129543e-01 -2.38922462e-01
8.92726406e-02 5.26396573e-01 1.38070893e+00 -9.59869385e-01
8.51157382e-02 -4.64300305e-01 -3.73193949e-01 3.59718829e-01
9.35888216e-02 -1.14319054e-02 -3.49452913e-01 -6.03750527e-01
-8.14210027e-02 -6.17185161e-02 3.68475854e-01 8.44006658e-01
1.43769395e+00 -2.23402396e-01 -1.40868798e-01 4.94632959e-01
1.38237643e+00 1.49792463e-01 5.18607616e-01 -2.48582177e-02
2.42452193e-02 4.29347426e-01 9.58208263e-01 7.27666795e-01
1.70527250e-01 3.27766716e-01 -1.62437961e-01 1.23428144e-01
5.42570829e-01 2.36851256e-02 5.07832587e-01 6.93780899e-01
2.93819278e-01 1.40286121e-03 -7.61146128e-01 4.73781914e-01
-1.67929089e+00 -1.06534839e+00 -3.72565925e-01 3.31970549e+00
1.02081645e+00 2.34524295e-01 2.96846509e-01 1.42191380e-01
9.30941999e-01 -4.19354111e-01 -5.64960837e-01 -6.13063991e-01
-3.86064976e-01 7.65503824e-01 8.00145149e-01 5.90817094e-01
-7.97787547e-01 1.89791948e-01 7.29334688e+00 1.21655846e+00
-6.50862575e-01 -8.62279162e-02 6.18229628e-01 9.49939787e-02
-4.13070858e-01 5.31533025e-02 -6.90088749e-01 6.06225371e-01
9.68587339e-01 -6.21039152e-01 -4.85172532e-02 8.42025042e-01
4.48001891e-01 -8.04470718e-01 -1.19700909e+00 8.48289847e-01
-8.62357318e-02 -6.98771417e-01 -4.13105696e-01 5.57856262e-01
7.42039800e-01 -4.88801062e-01 3.27756852e-01 3.68296146e-01
1.34513259e-01 -1.01361907e+00 3.49051595e-01 2.43948236e-01
1.34645164e+00 -1.37682939e+00 9.96760011e-01 2.93080658e-01
-1.01893687e+00 2.71749794e-01 -6.42648995e-01 9.68854055e-02
-3.52304071e-01 1.23859477e+00 -6.39861345e-01 2.18210667e-01
5.03296912e-01 -9.22803134e-02 -3.33745599e-01 1.31423223e+00
2.58548349e-01 6.78396523e-01 -4.50404942e-01 -3.77036482e-02
-6.63553132e-03 -5.96986592e-01 4.86730427e-01 1.14799833e+00
9.48931456e-01 -7.27823451e-02 2.25936800e-01 6.62817776e-01
8.33721757e-02 2.46307224e-01 -7.29872167e-01 1.48387998e-01
7.90295005e-01 6.74301863e-01 -7.83994436e-01 -1.57003939e-01
-3.90484035e-01 8.50367785e-01 5.39345108e-02 3.17002743e-01
-8.60022843e-01 -1.00164974e+00 8.30548406e-01 2.08937764e-01
2.80242085e-01 -2.05756605e-01 -4.77562845e-01 -8.81740630e-01
4.95566539e-02 -6.96505725e-01 5.23990870e-01 -2.81767845e-01
-1.51479042e+00 -2.43577272e-01 3.58812034e-01 -1.27017307e+00
-2.75269628e-01 -8.80691230e-01 -5.49768865e-01 1.03791475e+00
-7.98813462e-01 -1.99834138e-01 2.98972130e-01 5.27288139e-01
-2.75462598e-01 2.53949642e-01 7.32708275e-01 1.19311385e-01
-4.11371976e-01 1.15959048e+00 6.31775439e-01 -1.14780806e-01
1.09844160e+00 -1.61163211e+00 -1.63679361e-01 7.47646391e-01
-6.88757718e-01 6.60013437e-01 9.19832051e-01 -8.89473796e-01
-1.49300492e+00 -8.43294144e-01 8.18990290e-01 -1.51268035e-01
8.88989687e-01 -2.36450687e-01 -9.62376773e-01 3.89281005e-01
-3.48163933e-01 1.01566516e-01 1.22801709e+00 3.22137773e-01
-2.95381486e-01 -3.17188829e-01 -1.55639625e+00 6.96284652e-01
6.65298462e-01 -3.50280225e-01 -4.03960973e-01 1.24179780e-01
2.69539028e-01 1.80610642e-01 -1.12897933e+00 7.04534292e-01
7.47858644e-01 -1.14334702e+00 5.28860211e-01 -4.96601522e-01
-1.66819260e-01 -4.22178686e-01 -2.42012367e-01 -1.08831036e+00
-4.96148884e-01 -7.27293313e-01 2.68587023e-01 9.73772228e-01
4.27379996e-01 -1.04323721e+00 6.24352813e-01 3.06396544e-01
1.88724235e-01 -8.28379750e-01 -1.45443916e+00 -1.28817642e+00
2.53517240e-01 -6.18046641e-01 3.02104592e-01 9.38226938e-01
2.79876441e-01 -1.26764998e-01 -1.57160014e-01 2.73527831e-01
1.13259494e+00 1.09774537e-01 7.40941405e-01 -1.36663568e+00
-5.53133607e-01 -5.18612504e-01 -6.32152259e-01 -9.08039093e-01
1.43700704e-01 -7.86058843e-01 1.75736144e-01 -9.25916016e-01
3.61520648e-01 -2.49695092e-01 -2.42560476e-01 1.68600276e-01
-3.51939946e-01 1.85319275e-01 -7.56181002e-01 -3.43062282e-01
-2.79420108e-01 2.98128307e-01 1.12668455e+00 2.63852805e-01
-3.42928052e-01 2.19293743e-01 -6.12895489e-01 5.84423423e-01
8.13023090e-01 -4.78981376e-01 -3.68248284e-01 2.88825929e-01
2.83890635e-01 2.98287213e-01 8.31599310e-02 -8.32903445e-01
5.38564026e-02 -4.39044029e-01 4.87810761e-01 -7.51986742e-01
-2.24829704e-01 -2.69120753e-01 2.50017755e-02 6.06243134e-01
-2.37479266e-02 2.99437523e-01 1.87710091e-01 5.88775039e-01
3.48361358e-02 -2.92550296e-01 1.05751646e+00 5.04368007e-01
3.93429212e-02 5.29192686e-02 -4.76375937e-01 4.24196810e-01
1.09871733e+00 -4.15345252e-01 -3.37093711e-01 -1.92439854e-01
-3.25885981e-01 1.49719000e-01 4.52592015e-01 -3.52414250e-01
4.99623090e-01 -1.55605352e+00 -7.91833758e-01 3.85463089e-01
1.10669896e-01 -2.83650279e-01 1.29550859e-01 1.43593311e+00
-4.36842203e-01 3.00081879e-01 -2.60634068e-02 -6.27401352e-01
-9.92288709e-01 3.10191900e-01 -5.87334223e-02 2.24554166e-02
-1.12174943e-01 8.06473792e-01 3.80540341e-01 -3.07622164e-01
-1.46624207e-01 -2.67361194e-01 4.74328876e-01 -4.67011966e-02
6.49699569e-01 9.23717499e-01 -9.71543565e-02 -9.49022844e-02
-4.78532374e-01 4.19042617e-01 3.29284370e-01 -5.63360274e-01
9.80665505e-01 1.12969711e-01 -6.32895172e-01 5.73495805e-01
1.35118282e+00 6.07976437e-01 -9.01926577e-01 1.14661448e-01
8.32078755e-02 -8.80852640e-01 -4.27039534e-01 -2.93402940e-01
-2.64663070e-01 3.34464550e-01 3.95228475e-01 4.24701244e-01
9.89893913e-01 -1.27861574e-01 8.54617208e-02 3.10821265e-01
4.62950528e-01 -1.01897502e+00 -3.01281869e-01 4.79766488e-01
4.08949763e-01 -8.99064124e-01 5.38438819e-02 -3.99544030e-01
-6.52867183e-02 1.01368582e+00 5.12797356e-01 -2.68635362e-01
8.38842511e-01 3.50227982e-01 -4.81041938e-01 2.65557349e-01
-3.39757264e-01 -5.45580965e-03 4.05485749e-01 6.86137438e-01
4.41145182e-01 3.44800562e-01 -8.63415062e-01 7.51069486e-01
-2.61578143e-01 -2.39369735e-01 4.50508982e-01 8.52875948e-01
-7.08919287e-01 -9.58405554e-01 -4.91584420e-01 9.56967652e-01
-3.46859455e-01 4.16459981e-03 7.64839798e-02 1.04080224e+00
-6.69778436e-02 9.42420840e-01 2.91049480e-01 -2.40268543e-01
1.56320691e-01 1.03111550e-01 6.91018999e-01 -3.68418157e-01
3.04722190e-01 2.65698940e-01 -2.79890180e-01 -6.66277111e-01
4.49177511e-02 -6.65872574e-01 -9.10408795e-01 -8.64326596e-01
-3.15713137e-01 6.12602651e-01 1.75985023e-01 8.36952031e-01
2.03317270e-01 -1.59695536e-01 8.90276670e-01 -2.72265524e-01
-1.01692104e+00 -8.37978542e-01 -1.27952135e+00 -1.58118963e-01
4.17023040e-02 -7.26412356e-01 -7.90651441e-01 -6.46659255e-01]
|
[7.476391315460205, 4.376537799835205]
|
5acadc90-1ca6-4cc6-9ba9-a65e96163024
|
image-cropping-with-composition-and-saliency
|
1911.10492
| null |
https://arxiv.org/abs/1911.10492v1
|
https://arxiv.org/pdf/1911.10492v1.pdf
|
Image Cropping with Composition and Saliency Aware Aesthetic Score Map
|
Aesthetic image cropping is a practical but challenging task which aims at finding the best crops with the highest aesthetic quality in an image. Recently, many deep learning methods have been proposed to address this problem, but they did not reveal the intrinsic mechanism of aesthetic evaluation. In this paper, we propose an interpretable image cropping model to unveil the mystery. For each image, we use a fully convolutional network to produce an aesthetic score map, which is shared among all candidate crops during crop-level aesthetic evaluation. Then, we require the aesthetic score map to be both composition-aware and saliency-aware. In particular, the same region is assigned with different aesthetic scores based on its relative positions in different crops. Moreover, a visually salient region is supposed to have more sensitive aesthetic scores so that our network can learn to place salient objects at more proper positions. Such an aesthetic score map can be used to localize aesthetically important regions in an image, which sheds light on the composition rules learned by our model. We show the competitive performance of our model in the image cropping task on several benchmark datasets, and also demonstrate its generality in real-world applications.
|
['Weijie Zhao', 'Yi Tu', 'Li Niu', 'Dawei Cheng', 'Liqing Zhang']
|
2019-11-24
| null | null | null | null |
['image-cropping']
|
['computer-vision']
|
[ 3.22055608e-01 1.40866667e-01 -1.18498348e-01 -3.78014982e-01
-2.69413054e-01 -4.39249694e-01 1.23301759e-01 3.65737438e-01
1.54345647e-01 1.88394353e-01 2.39565045e-01 -4.27032337e-02
2.65402999e-02 -9.24598455e-01 -8.61595333e-01 -7.64178634e-01
1.78139478e-01 -1.00836396e-01 3.20286244e-01 -4.43170965e-01
3.69159192e-01 1.28855333e-01 -1.67245674e+00 3.50694925e-01
1.09061086e+00 1.17476952e+00 7.31589735e-01 2.32250482e-01
-8.19579437e-02 5.41554809e-01 -6.89149439e-01 -4.26065475e-01
2.39753887e-01 -5.07887423e-01 -5.77454031e-01 3.84224176e-01
1.02432504e-01 -1.09484486e-01 2.76083916e-01 1.40380263e+00
6.67206198e-02 -1.69367939e-01 6.78558886e-01 -1.26641798e+00
-1.25221169e+00 6.92112446e-01 -1.01800847e+00 -2.09652349e-01
-1.05372265e-01 2.08384633e-01 1.45127273e+00 -7.78015673e-01
3.64840955e-01 1.23826683e+00 6.93034232e-02 2.73988873e-01
-9.92329717e-01 -5.11880934e-01 3.84941965e-01 2.95732796e-01
-9.52847600e-01 1.10509627e-01 1.19011116e+00 -1.84592858e-01
-4.91536744e-02 4.76991951e-01 1.03018641e+00 5.44857204e-01
2.09216829e-02 1.26449502e+00 1.10433340e+00 -3.53569478e-01
3.36120605e-01 1.68104365e-01 -2.98678696e-01 4.95929003e-01
2.66706556e-01 -2.23558754e-01 -5.45307636e-01 3.76133591e-01
7.88847387e-01 1.20165594e-01 -2.45381460e-01 -6.45757496e-01
-1.26988661e+00 7.92750239e-01 1.24212909e+00 3.09522659e-01
-4.19079125e-01 1.77299947e-01 2.50839680e-01 -7.92849734e-02
3.64848644e-01 9.73093987e-01 -4.84130830e-01 5.53710938e-01
-5.82899928e-01 5.96631356e-02 4.25732732e-02 7.22066224e-01
1.12834179e+00 -2.63891250e-01 -3.92904818e-01 7.97000706e-01
2.51119494e-01 4.02327955e-01 1.56083956e-01 -9.66966033e-01
1.69540733e-01 9.97951448e-01 9.70391110e-02 -1.48392010e+00
-1.90928265e-01 -5.10923326e-01 -1.01931059e+00 4.03606385e-01
1.90853700e-01 1.61696538e-01 -7.07857609e-01 1.82054889e+00
2.12930843e-01 -2.00912997e-01 1.16494164e-01 1.27004278e+00
7.39635170e-01 6.96540415e-01 2.02340752e-01 1.72976121e-01
1.65705562e+00 -1.09224570e+00 -5.97077966e-01 -4.70386595e-01
2.21008763e-01 -8.99040818e-01 1.61860001e+00 2.30221763e-01
-9.33026314e-01 -4.70877856e-01 -1.14997160e+00 -1.26724660e-01
-3.92388701e-01 4.97745574e-01 9.04936433e-01 3.27457547e-01
-9.37649608e-01 4.33490306e-01 -2.14701951e-01 -2.79528737e-01
5.64121187e-01 -1.47765577e-01 -2.46716049e-02 2.04365969e-01
-1.15110099e+00 6.97110415e-01 5.19626498e-01 2.78514266e-01
-7.42179215e-01 -3.48942131e-01 -8.68487120e-01 4.21643674e-01
3.31526488e-01 -3.49988967e-01 9.29334700e-01 -1.65679359e+00
-1.01701164e+00 9.14012849e-01 -1.38436303e-01 -2.62823645e-02
6.87173828e-02 6.32520244e-02 -1.81336418e-01 2.38792360e-01
3.94832164e-01 1.07584846e+00 9.00056958e-01 -1.71016371e+00
-7.48849154e-01 -2.07012713e-01 4.03246015e-01 4.09938097e-01
-6.46363318e-01 -1.19899824e-01 -4.56947386e-01 -8.19443285e-01
1.94373712e-01 -5.23405015e-01 -2.17111737e-01 3.58914554e-01
-5.97750068e-01 -1.44694686e-01 5.77700734e-01 -5.00946343e-01
1.12156630e+00 -2.04326916e+00 1.14691831e-01 1.65814191e-01
2.42158696e-01 2.36434400e-01 -2.84196377e-01 8.54538530e-02
-1.04989201e-01 3.12450409e-01 -3.67752761e-01 1.42405748e-01
-1.06345557e-01 3.24634761e-02 -2.85084933e-01 -1.28173605e-01
7.44507790e-01 1.34311199e+00 -1.15568328e+00 -5.53647935e-01
1.34782255e-01 9.69835222e-02 -3.84039283e-01 1.89735636e-01
-4.26102459e-01 3.12987775e-01 -7.21053660e-01 8.30575585e-01
8.88131618e-01 -4.30466145e-01 1.99505873e-02 -4.34875995e-01
-4.13935296e-02 -2.38876894e-01 -7.77310669e-01 1.55994761e+00
-3.87189209e-01 7.01138914e-01 -2.30398998e-01 -9.23215330e-01
1.23679769e+00 -3.10496122e-01 2.16828331e-01 -9.98683870e-01
4.79646176e-02 2.55390823e-01 -2.12789848e-01 -5.76708972e-01
7.49658763e-01 -4.51410674e-02 -1.99403822e-01 4.89799291e-01
-4.57296401e-01 -1.38531506e-01 -9.53738019e-02 6.75883889e-02
5.21245718e-01 4.32415828e-02 2.20901281e-01 -4.65046376e-01
4.39143747e-01 -5.83264865e-02 5.20761132e-01 2.59279639e-01
-2.67661572e-01 8.10335994e-01 6.90577626e-01 -6.49697602e-01
-1.08642018e+00 -1.01155078e+00 1.20992310e-01 1.19795334e+00
9.84256685e-01 -6.86627105e-02 -9.24987912e-01 -6.14295065e-01
-2.56168872e-01 6.17345572e-01 -1.05383730e+00 -5.40808201e-01
-4.76720273e-01 -7.70492315e-01 -4.18111756e-02 3.42847079e-01
6.58494353e-01 -1.68921459e+00 -9.68123496e-01 -3.16309668e-02
-5.57177842e-01 -6.06709838e-01 -6.41591072e-01 2.02296481e-01
-5.57474971e-01 -1.18055642e+00 -1.05309439e+00 -1.14019310e+00
1.08465767e+00 8.41271877e-01 1.32773149e+00 3.18698227e-01
-6.38994351e-02 -2.84216017e-01 -7.04553246e-01 -6.02297902e-01
-2.30665565e-01 1.46737605e-01 -6.00043356e-01 3.25727940e-01
2.41107360e-01 -2.93585569e-01 -1.04705942e+00 4.54902589e-01
-1.25553787e+00 4.08566117e-01 9.25675452e-01 8.27565312e-01
6.87335193e-01 1.32949486e-01 5.68477809e-01 -5.44526756e-01
7.50786364e-01 -2.91451931e-01 -4.11912411e-01 6.43422484e-01
-4.11726028e-01 3.57900113e-01 4.14181381e-01 -3.12899500e-01
-8.27231884e-01 1.66092381e-01 7.67944604e-02 -3.13016355e-01
7.26781264e-02 4.78502691e-01 -4.89544749e-01 1.03479708e-02
4.33622599e-01 3.55310321e-01 -9.47836190e-02 -2.55963027e-01
5.80808222e-01 3.65423054e-01 3.60902488e-01 -4.59319174e-01
7.12106287e-01 4.04280305e-01 -6.67821169e-02 -2.31221467e-01
-1.13547790e+00 -7.12379143e-02 -5.00670314e-01 -2.98949957e-01
8.96651030e-01 -6.54028058e-01 -6.78393841e-01 4.25231934e-01
-1.32350528e+00 -1.36837646e-01 -4.07409668e-01 -2.50046700e-01
-4.25216287e-01 3.26403201e-01 1.67914052e-02 -7.64539957e-01
-4.41186368e-01 -1.23871410e+00 1.44734716e+00 7.02075660e-01
7.11858571e-02 -5.82078993e-01 -3.27084303e-01 -1.50011212e-01
3.30469638e-01 4.03703660e-01 1.15615249e+00 1.30238697e-01
-7.73607671e-01 1.14291951e-01 -8.96010578e-01 1.42077789e-01
2.28968605e-01 1.95368066e-01 -9.74307358e-01 1.49208620e-01
-3.19095939e-01 -2.04875648e-01 1.05050671e+00 5.36949396e-01
1.54766428e+00 -3.82494837e-01 -1.67136922e-01 5.03512263e-01
1.37453735e+00 1.64039195e-01 8.24342132e-01 4.87868935e-01
6.61771059e-01 9.86265659e-01 9.61353242e-01 2.55028278e-01
3.47229272e-01 5.72899103e-01 1.18295836e+00 -8.09213340e-01
1.86049156e-02 -4.97674912e-01 1.17803730e-01 3.15514922e-01
2.02665091e-01 -2.01117977e-01 -4.86706495e-01 7.17279375e-01
-1.96762478e+00 -7.57930160e-01 -2.52659153e-02 2.08141208e+00
7.23615766e-01 7.33512640e-02 9.06154960e-02 5.31094484e-02
1.01526070e+00 3.18635970e-01 -8.11841011e-01 -3.16558778e-01
-4.74152625e-01 -1.22559696e-01 2.99346894e-01 7.53307194e-02
-1.18587732e+00 1.13002169e+00 5.91584826e+00 9.58968282e-01
-1.14833677e+00 -1.54290929e-01 1.25010049e+00 2.99547344e-01
-7.57800817e-01 7.93671533e-02 -2.12610409e-01 5.27796745e-01
-1.28438041e-01 -2.17444688e-01 3.58793288e-01 8.73064220e-01
1.73869178e-01 -2.60222971e-01 -5.24530590e-01 7.40049124e-01
-4.34880592e-02 -1.18246353e+00 2.68642157e-01 -8.21397081e-02
7.33213365e-01 -5.96835613e-01 4.87956285e-01 -1.16995551e-01
1.58752471e-01 -1.00045562e+00 1.00614667e+00 4.17009145e-01
7.29142964e-01 -8.57655108e-01 5.58803141e-01 1.37489930e-01
-1.50987077e+00 -2.14461505e-01 -8.06715369e-01 4.43494730e-02
-1.97515786e-02 8.24329853e-01 -5.11615217e-01 4.07400548e-01
7.83558607e-01 8.80251586e-01 -8.92712295e-01 1.02710414e+00
-5.63692629e-01 1.63326472e-01 7.68235102e-02 -2.67317653e-01
3.02694023e-01 -2.50205338e-01 9.66261700e-02 7.63274193e-01
5.09455979e-01 -1.84200466e-01 -1.33076891e-01 1.29751408e+00
-9.03788880e-02 2.98794389e-01 -4.31847423e-01 1.10294878e-01
2.02386901e-01 1.67070675e+00 -1.06612861e+00 -1.13667972e-01
2.50117201e-02 1.13057125e+00 3.83333623e-01 2.18974948e-01
-8.69623840e-01 -3.74023885e-01 6.93836927e-01 -1.01381615e-02
2.84306854e-01 1.96602583e-01 -6.14800215e-01 -1.05287349e+00
1.06889419e-01 -5.88879108e-01 1.34348899e-01 -1.28341711e+00
-1.18858576e+00 6.21109009e-01 -4.27516222e-01 -1.56231773e+00
3.34027499e-01 -5.38787484e-01 -9.71578896e-01 8.09398174e-01
-1.81620800e+00 -1.20017886e+00 -7.04074204e-01 1.92380875e-01
5.69649279e-01 1.92271039e-01 5.88926256e-01 -7.39400163e-02
-3.66224051e-01 4.54014987e-01 -1.25584528e-01 -3.36735360e-02
6.22933447e-01 -1.33664691e+00 5.15261412e-01 8.20600748e-01
1.01509638e-01 3.30137342e-01 7.17060804e-01 -3.76921386e-01
-8.06377590e-01 -1.21063137e+00 8.30056608e-01 -1.83343440e-01
3.46198052e-01 -2.44620115e-01 -7.96476245e-01 -3.29693519e-02
3.64252895e-01 -2.25174904e-01 1.84051752e-01 -3.38428542e-02
-4.27335769e-01 -3.18688869e-01 -9.24345493e-01 8.10246110e-01
9.99938786e-01 -1.27774179e-01 -3.22635770e-01 3.42470855e-01
1.11115015e+00 -4.46338989e-02 -3.14878881e-01 3.27100366e-01
5.17560840e-01 -1.24150002e+00 1.05021870e+00 -2.72951335e-01
1.01819825e+00 -5.44734776e-01 1.25112161e-01 -1.45628667e+00
-5.67133307e-01 -2.10235029e-01 3.25733513e-01 1.10863864e+00
3.98667812e-01 -2.74580449e-01 5.19897997e-01 1.99678779e-01
-3.46839353e-02 -1.00559008e+00 -3.97268653e-01 -3.52225989e-01
-1.08861461e-01 -1.06587976e-01 1.15677917e+00 7.52477348e-01
-1.91663578e-01 2.16990009e-01 -5.29188097e-01 1.26451924e-01
3.43347996e-01 7.95381248e-01 5.59848666e-01 -1.23278725e+00
4.17396724e-02 -8.61656189e-01 -2.55080372e-01 -9.32147503e-01
-1.37807250e-01 -6.87197924e-01 2.57696331e-01 -1.56345093e+00
5.91520905e-01 -6.06730580e-01 -5.92916965e-01 7.31510103e-01
-7.34943986e-01 4.39534843e-01 3.09637159e-01 2.40658700e-01
-8.07022810e-01 7.70223856e-01 1.81585205e+00 -4.19479966e-01
-3.76495235e-02 -1.67627037e-01 -1.49318171e+00 5.15401721e-01
9.74238336e-01 -1.31640658e-01 -4.05460626e-01 -4.99461025e-01
5.59238255e-01 -4.83797133e-01 6.13854229e-01 -7.34335601e-01
-2.31650323e-01 -5.21437883e-01 5.65999150e-01 -5.22429287e-01
6.65730312e-02 -8.26071203e-01 -1.32266387e-01 4.67483193e-01
-4.44788784e-01 4.09439616e-02 -6.18530959e-02 3.84898156e-01
-2.08135024e-01 -1.78036124e-01 7.91135967e-01 -1.04162104e-01
-1.02993309e+00 2.25845739e-01 1.97656006e-01 -1.10767253e-01
1.13817477e+00 -1.27680704e-01 -3.15491170e-01 -4.55523968e-01
-2.58789361e-01 3.17298949e-01 7.83656061e-01 6.36741877e-01
7.22309768e-01 -1.49338400e+00 -5.67464292e-01 3.25365961e-01
6.50445879e-01 4.94513512e-02 3.37846339e-01 4.48188186e-01
-4.03930724e-01 2.99229054e-03 -5.98380148e-01 -6.05306327e-01
-1.04072618e+00 9.44223940e-01 2.10559160e-01 -1.57932594e-01
-3.42869163e-01 7.91336060e-01 9.90855575e-01 -1.62058085e-01
-6.16067797e-02 -3.15848827e-01 -5.85463822e-01 7.03766197e-02
4.53322858e-01 -1.46585122e-01 -2.87418246e-01 -5.29268920e-01
-1.80997789e-01 9.37911093e-01 1.60229638e-01 1.76116392e-01
1.19310558e+00 -2.93550014e-01 -1.56689897e-01 1.93150595e-01
9.71102357e-01 -2.15475500e-01 -1.40889525e+00 -1.66557729e-01
-9.78345200e-02 -6.91078067e-01 -1.20734319e-01 -7.96935737e-01
-1.63697660e+00 1.04831970e+00 5.06847441e-01 4.96087343e-01
1.72801054e+00 1.86421707e-01 7.40528882e-01 5.71519323e-03
2.51126617e-01 -1.08129323e+00 5.28716862e-01 6.26476854e-02
1.12443554e+00 -1.51256466e+00 -1.76021770e-01 -4.91113901e-01
-1.11910868e+00 1.02798355e+00 6.81391060e-01 -1.45078972e-01
3.18852603e-01 -1.87479317e-01 1.89008728e-01 -3.97857159e-01
-2.20933616e-01 -5.31602919e-01 4.73773897e-01 6.86574638e-01
2.53739953e-01 4.47227597e-01 -2.72409528e-01 7.13695109e-01
-7.24956170e-02 -3.56319934e-01 2.36882254e-01 4.37301934e-01
-7.95412719e-01 -8.78934145e-01 -3.12965721e-01 2.19749197e-01
1.09863251e-01 -1.20369107e-01 -6.14964902e-01 2.92513371e-01
3.65891695e-01 8.40360463e-01 -1.69586949e-02 -5.16438961e-01
2.08416894e-01 -4.66008037e-01 1.33584201e-01 -3.15915167e-01
-3.73539090e-01 1.02637216e-01 -5.35220385e-01 -4.60593045e-01
-4.42564398e-01 -4.36716318e-01 -1.17026460e+00 -1.14353634e-01
-2.27189183e-01 -9.87578835e-03 5.09305477e-01 5.77401638e-01
4.24765408e-01 7.24779665e-01 9.92996037e-01 -8.45640719e-01
1.05218299e-01 -6.02425396e-01 -7.65888691e-01 5.03206432e-01
1.84633225e-01 -6.89876735e-01 -8.26915577e-02 6.61365762e-02]
|
[11.496810913085938, -0.9444612860679626]
|
6ab175a7-1e76-48cb-a979-1c2a74fb84b6
|
self-supervised-optimization-of-hand-pose
|
2307.03007
| null |
https://arxiv.org/abs/2307.03007v1
|
https://arxiv.org/pdf/2307.03007v1.pdf
|
Self-supervised Optimization of Hand Pose Estimation using Anatomical Features and Iterative Learning
|
Manual assembly workers face increasing complexity in their work. Human-centered assistance systems could help, but object recognition as an enabling technology hinders sophisticated human-centered design of these systems. At the same time, activity recognition based on hand poses suffers from poor pose estimation in complex usage scenarios, such as wearing gloves. This paper presents a self-supervised pipeline for adapting hand pose estimation to specific use cases with minimal human interaction. This enables cheap and robust hand posebased activity recognition. The pipeline consists of a general machine learning model for hand pose estimation trained on a generalized dataset, spatial and temporal filtering to account for anatomical constraints of the hand, and a retraining step to improve the model. Different parameter combinations are evaluated on a publicly available and annotated dataset. The best parameter and model combination is then applied to unlabelled videos from a manual assembly scenario. The effectiveness of the pipeline is demonstrated by training an activity recognition as a downstream task in the manual assembly scenario.
|
['Marco F. Huber', 'Timo Leitritz', 'Christian Jauch']
|
2023-07-06
| null | null | null | null |
['pose-estimation', 'activity-recognition', 'object-recognition', 'hand-pose-estimation']
|
['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision']
|
[ 3.75032037e-01 9.12148505e-02 -2.80502707e-01 -2.15358049e-01
-6.36553645e-01 -6.70803428e-01 1.72741473e-01 -2.27654606e-01
-6.72970355e-01 4.52433079e-01 2.38515303e-01 -8.51036906e-02
-3.18350732e-01 1.12466224e-01 -2.87382185e-01 -6.33307993e-01
2.02272236e-01 8.52176547e-01 3.29597592e-01 -4.08467166e-02
3.69137704e-01 1.05549824e+00 -1.56009352e+00 3.87668639e-01
3.14321846e-01 6.20414197e-01 5.67284048e-01 9.28054333e-01
2.84984589e-01 1.21307962e-01 -7.68297970e-01 -1.85743451e-01
3.20098251e-01 -3.70414615e-01 -6.71899676e-01 5.58740973e-01
5.43130398e-01 -5.88922501e-01 -1.66574150e-01 1.20199323e-01
1.15788209e+00 1.68247819e-01 4.61094558e-01 -8.50304842e-01
3.10484082e-01 -8.97890106e-02 -2.95083106e-01 2.37863719e-01
5.85889578e-01 6.50312185e-01 3.01035494e-01 -7.44650722e-01
8.40890765e-01 8.97892654e-01 7.63182521e-01 6.95014894e-01
-1.14663553e+00 -2.68583089e-01 1.15421809e-01 1.32420555e-01
-1.32812059e+00 -4.75086004e-01 6.02770805e-01 -8.81993711e-01
1.41202819e+00 2.34204918e-01 8.07947874e-01 1.27493322e+00
1.83399260e-01 5.50768852e-01 9.27576542e-01 -5.30958354e-01
2.09902182e-01 2.10643381e-01 -1.99460745e-01 6.13251925e-01
2.98753411e-01 -2.10439786e-01 -8.01956356e-01 2.22258223e-03
8.46994281e-01 1.26785174e-01 -1.31814688e-01 -6.21451855e-01
-1.25165486e+00 1.59560397e-01 2.51484692e-01 5.65696299e-01
-6.00914061e-01 -8.43903869e-02 3.12586844e-01 -1.87385067e-01
2.37792701e-01 4.95584637e-01 -7.92305470e-01 -5.11908591e-01
-1.05108643e+00 3.27894837e-01 8.68123293e-01 9.22207713e-01
1.40899479e-01 -3.70905370e-01 -7.45206714e-01 5.39921463e-01
3.45053971e-01 2.21869737e-01 2.86933035e-01 -6.63151205e-01
5.00832915e-01 6.84308887e-01 2.46672645e-01 -3.35252166e-01
-9.19647038e-01 -5.98734140e-01 -3.43289860e-02 5.08590937e-01
7.74921834e-01 -6.76464364e-02 -1.03884888e+00 1.00347948e+00
6.23186827e-01 -4.72781479e-01 -6.13149405e-01 1.21095729e+00
2.27612838e-01 -2.35943630e-01 1.96747243e-01 -2.38501012e-01
1.39419460e+00 -8.90234470e-01 -7.75176287e-01 -4.76529390e-01
7.73511171e-01 -7.91347861e-01 1.08927608e+00 8.42923403e-01
-7.95752704e-01 -5.58983922e-01 -1.09585118e+00 -1.31515741e-01
-2.87404239e-01 7.89420187e-01 6.11511528e-01 8.89433563e-01
-6.01389110e-01 8.42925549e-01 -1.36078811e+00 -6.41290069e-01
4.02792454e-01 8.52745831e-01 -5.93401611e-01 1.73862010e-01
-3.53853613e-01 1.23896658e+00 4.14274305e-01 5.37866235e-01
-5.12213111e-01 -3.82387549e-01 -7.40910351e-01 -3.43126684e-01
6.54406965e-01 -5.75323582e-01 1.33537781e+00 -4.40130323e-01
-1.66302776e+00 8.88124108e-01 -2.19996303e-01 -8.76080021e-02
1.05827045e+00 -5.80670416e-01 -2.26941742e-02 7.30923414e-02
-1.80093318e-01 7.34312907e-02 1.07496607e+00 -9.28831756e-01
-3.83090436e-01 -8.75397503e-01 -2.09642828e-01 3.88649374e-01
-1.84006646e-01 7.09319338e-02 -7.36035466e-01 -5.93722522e-01
7.49181062e-02 -1.07581055e+00 -8.79513249e-02 2.23472551e-01
1.56645235e-02 -3.70230004e-02 8.17160308e-01 -1.37688160e+00
1.12595856e+00 -2.05353212e+00 3.89681995e-01 2.63311982e-01
-1.39972433e-01 2.59875506e-01 1.42061561e-01 2.83908695e-01
8.58734250e-02 -5.62462747e-01 -2.34208349e-02 -4.96803671e-01
-1.17684156e-01 8.95610377e-02 4.78277355e-01 6.48203015e-01
1.23247243e-01 6.58120453e-01 -7.10955262e-01 -6.13685429e-01
5.85005164e-01 7.02935696e-01 -5.09208858e-01 3.65973562e-01
-2.99891736e-02 1.19405329e+00 -1.81357916e-02 7.35232770e-01
9.81978700e-02 3.03337891e-02 4.62814808e-01 -3.68312597e-01
1.18164882e-01 7.92932734e-02 -1.26320362e+00 2.26838470e+00
-6.18025243e-01 2.03569174e-01 3.19122165e-01 -3.98186833e-01
5.07085204e-01 4.96222585e-01 6.00969911e-01 -1.02414601e-01
5.31603813e-01 2.60582983e-01 2.42163867e-01 -8.97023916e-01
4.69937688e-03 4.76827733e-02 4.28400010e-01 3.23928356e-01
3.97132725e-01 -1.11049645e-01 2.15144753e-01 -2.75101095e-01
1.05261004e+00 8.59126627e-01 3.73041213e-01 2.89171897e-02
3.11171889e-01 -1.13802910e-01 -2.42304504e-02 4.40433115e-01
-2.61108339e-01 7.94420838e-01 -1.59591492e-02 -4.32305723e-01
-8.33741486e-01 -7.21204937e-01 2.86186546e-01 1.20710278e+00
-3.32485318e-01 -4.41210330e-01 -8.80269170e-01 -1.04056680e+00
6.76772604e-03 4.03967619e-01 -5.41758180e-01 -2.63855383e-02
-7.55750835e-01 -2.92260826e-01 1.45292819e-01 1.02318335e+00
9.57894698e-02 -1.13559270e+00 -1.07707798e+00 1.66924804e-01
9.70847458e-02 -9.40003216e-01 -5.34379840e-01 2.63631940e-01
-8.29232275e-01 -1.24446523e+00 -8.96113753e-01 -6.15109444e-01
7.02190220e-01 -2.62325585e-01 5.62161565e-01 -8.07397664e-02
-8.45212936e-01 7.19873667e-01 -1.10565893e-01 -5.34222007e-01
-9.66195986e-02 4.04001147e-01 2.77538300e-01 1.48600293e-02
1.12977184e-01 -3.77894491e-01 -7.32555270e-01 4.21763837e-01
-2.91005820e-01 -3.00563067e-01 9.55928385e-01 7.60576367e-01
4.64962721e-01 -5.02011180e-01 1.38556911e-02 -5.81186652e-01
5.64266384e-01 2.57246107e-01 -2.61577398e-01 4.33659345e-01
-5.99895120e-01 -3.84809412e-02 -5.79606667e-02 -8.41977000e-01
-1.42685807e+00 1.02552891e+00 -1.75639004e-01 -3.52175087e-01
-3.86036009e-01 7.33312592e-02 -5.35468638e-01 -1.38548315e-01
9.37330246e-01 -1.76060021e-01 1.71093404e-01 -7.63619661e-01
1.52837023e-01 9.93553579e-01 7.80293107e-01 -4.15030777e-01
5.10484636e-01 3.43248904e-01 -4.59576249e-02 -8.27068150e-01
-6.22558296e-01 -6.92010701e-01 -1.71112359e+00 -4.83685613e-01
7.10294247e-01 -3.94851685e-01 -8.69409561e-01 3.67735445e-01
-1.19666862e+00 -4.43303645e-01 -1.90874979e-01 8.79243374e-01
-8.12160194e-01 1.87494099e-01 1.38444630e-02 -1.26006293e+00
-3.47546995e-01 -9.72251952e-01 1.45652199e+00 -1.42131057e-02
-7.79543877e-01 -4.98588651e-01 -6.13000728e-02 9.07503605e-01
1.68988764e-01 1.62473515e-01 2.13431016e-01 -5.56258976e-01
-1.48186207e-01 -8.82120669e-01 3.90411049e-01 2.06778333e-01
2.55921930e-01 -4.57144052e-01 -1.06482339e+00 -3.78609478e-01
-6.06947131e-02 -1.21041387e-01 2.63365716e-01 2.13218063e-01
1.00207388e+00 -2.34648436e-01 -3.78655910e-01 3.00839722e-01
6.06519938e-01 1.37646019e-01 4.15589869e-01 1.90347165e-01
8.25429201e-01 8.92275631e-01 8.82827282e-01 3.26128960e-01
-3.34665269e-01 8.85898888e-01 2.56442487e-01 1.64326563e-01
-2.56063074e-01 3.64439078e-02 3.53260972e-02 1.86244532e-01
-8.27792466e-01 1.42598733e-01 -1.07217216e+00 3.10953081e-01
-1.69718945e+00 -6.37079954e-01 1.45299390e-01 2.33022261e+00
6.28496706e-01 9.07101855e-02 5.87890923e-01 6.03061616e-01
3.43270838e-01 -4.15927857e-01 -6.89960420e-01 -9.75080358e-04
4.72117305e-01 3.33499402e-01 5.71204960e-01 3.05271775e-01
-1.14549530e+00 6.70191884e-01 6.04108477e+00 2.45354995e-01
-9.99042451e-01 4.77628887e-01 -1.83352992e-01 -5.12551427e-01
8.16213071e-01 -3.02240103e-01 -7.52312422e-01 2.80528158e-01
5.12073338e-01 5.91539621e-01 3.11729699e-01 9.54150200e-01
2.61727810e-01 -2.81940252e-01 -1.26474941e+00 1.28812325e+00
3.60378236e-01 -6.96126819e-01 -4.74719405e-01 5.56913652e-02
1.63417950e-01 -4.49629247e-01 -1.79998726e-01 2.30178759e-02
-7.76892424e-01 -7.99172342e-01 4.90042210e-01 6.60452068e-01
7.39612997e-01 -5.07167757e-01 7.14642465e-01 5.71262300e-01
-9.99636948e-01 -3.53524983e-01 4.49338377e-01 -1.43286139e-01
2.53264427e-01 -9.27089825e-02 -1.36625850e+00 2.18301117e-01
5.82189798e-01 2.11549461e-01 -8.18413675e-01 1.04938352e+00
-3.12220901e-01 2.50626117e-01 -3.42318267e-01 1.92668676e-01
-4.73822773e-01 3.57981026e-01 5.30001223e-01 1.35467541e+00
1.13760367e-01 -7.25921169e-02 1.20835982e-01 3.27915519e-01
4.89252925e-01 4.56748419e-02 -3.78771514e-01 -2.32661650e-01
9.14687570e-03 1.31746221e+00 -9.86018360e-01 2.10670114e-01
-1.05188064e-01 1.55204284e+00 9.60632414e-02 2.35855371e-01
-2.98021793e-01 -3.54758978e-01 3.09359670e-01 6.88767135e-01
-8.55145603e-03 -5.46802223e-01 -2.22027078e-01 -8.04658651e-01
5.42536438e-01 -8.78067315e-01 4.20053303e-01 -5.97126901e-01
-8.60538065e-01 2.49275401e-01 1.99998900e-01 -1.21780872e+00
-6.76769555e-01 -1.11322999e+00 -1.85345516e-01 9.61002469e-01
-3.76672715e-01 -1.52383828e+00 -6.94204092e-01 4.11874592e-01
7.57626712e-01 -2.49726146e-01 9.06554878e-01 2.04179943e-01
-4.61074919e-01 5.90161204e-01 -6.26846135e-01 -1.24107890e-01
8.82649839e-01 -1.04569960e+00 4.48997058e-02 7.25976050e-01
5.91359697e-02 7.57471323e-01 7.56657541e-01 -8.91073883e-01
-1.53108358e+00 -6.72334492e-01 7.79618502e-01 -1.08699977e+00
1.81710348e-01 -5.57250023e-01 -6.18364036e-01 7.83210516e-01
-2.38234133e-01 2.34048385e-02 5.01057327e-01 8.00173283e-02
-3.43960337e-02 5.33850379e-02 -1.30309129e+00 3.30193192e-01
1.57004786e+00 -5.27160347e-01 -6.59095347e-01 7.47606754e-01
-9.68775973e-02 -8.81106377e-01 -9.03878927e-01 3.27747077e-01
1.21280146e+00 -4.80039209e-01 7.83375084e-01 -6.21169508e-01
-2.67729610e-01 -3.46524745e-01 3.09559762e-01 -8.76719594e-01
-1.76157430e-01 -8.46525431e-01 -6.04684412e-01 9.64212477e-01
1.27144426e-01 -2.90248364e-01 1.03845990e+00 7.15982914e-01
3.30500528e-02 -7.39406884e-01 -1.04076028e+00 -9.09692824e-01
-6.91871166e-01 -3.18104982e-01 1.36118352e-01 3.00418258e-01
2.87521034e-01 3.64462852e-01 -2.61252016e-01 1.19310006e-01
3.79007876e-01 -4.13045675e-01 9.16411698e-01 -1.30315208e+00
-5.82816303e-01 -2.18889117e-01 -6.86346650e-01 -4.45217848e-01
-7.82055035e-02 -4.14803624e-01 2.59049237e-01 -1.60115600e+00
7.71591589e-02 1.37120396e-01 1.99767217e-01 7.35375464e-01
1.50877163e-02 3.21462631e-01 1.25677302e-01 7.45752454e-02
-1.25952929e-01 -3.13395081e-04 1.08968234e+00 -5.31095415e-02
-6.54710829e-01 3.58878225e-01 -1.03779569e-01 6.18573844e-01
8.72347832e-01 -2.33171970e-01 -4.63654280e-01 -1.01062641e-01
-3.17780152e-02 -2.36067235e-01 4.91184384e-01 -1.19499242e+00
7.37294480e-02 1.21698081e-01 7.25449264e-01 -6.54997766e-01
5.73688090e-01 -1.04919541e+00 3.35299999e-01 8.10378373e-01
-6.56752735e-02 -2.01599836e-01 2.11354777e-01 3.12194914e-01
3.61230642e-01 -1.31827399e-01 5.53025544e-01 -1.42866597e-01
-3.47756058e-01 -1.83851086e-03 -1.73212439e-01 -4.77636993e-01
1.13516283e+00 -7.09511220e-01 1.70028418e-01 -3.25717516e-02
-1.40406752e+00 -4.34614755e-02 2.46774569e-01 5.94100893e-01
3.52154136e-01 -8.83142292e-01 -3.76950830e-01 3.68863523e-01
2.38251001e-01 6.62382916e-02 2.67549574e-01 1.45777977e+00
-5.00838220e-01 4.65253860e-01 -3.80825549e-01 -6.38444066e-01
-1.79148293e+00 5.25191963e-01 5.66501915e-01 -7.40226656e-02
-5.47821701e-01 7.74986744e-01 -3.53471726e-01 -4.76098925e-01
7.42233753e-01 -2.61039883e-01 -9.51684937e-02 4.47394773e-02
6.05197847e-01 8.34695876e-01 6.63597822e-01 -7.42877603e-01
-6.48796380e-01 5.69175661e-01 2.18530759e-01 -3.86711508e-01
1.15805233e+00 3.70039605e-02 3.72285396e-01 4.06613886e-01
6.53313577e-01 -1.43766731e-01 -1.49168611e+00 1.64537489e-01
8.07184577e-02 -5.97185254e-01 -1.54980451e-01 -1.18081164e+00
-7.39154160e-01 9.24080610e-01 1.19235063e+00 -7.05093801e-01
8.80398333e-01 1.90029889e-01 1.26505479e-01 5.66694140e-01
5.45572698e-01 -1.44417489e+00 1.59054950e-01 4.94065471e-02
1.40154684e+00 -1.15259433e+00 2.93380201e-01 -4.88392144e-01
-6.71853304e-01 1.01711738e+00 7.28683650e-01 3.73886764e-01
5.43425083e-01 2.89749622e-01 1.90913215e-01 -2.38782719e-01
-1.88123286e-02 -3.50568712e-01 7.57225752e-01 9.82542694e-01
5.83049893e-01 -1.79797739e-01 -7.55983651e-01 6.90253973e-01
-7.03256130e-02 3.37200284e-01 -2.16536537e-01 1.67671299e+00
9.03605223e-02 -1.00659180e+00 -5.43476582e-01 6.12530529e-01
-4.25843239e-01 7.08397806e-01 -7.09251821e-01 8.27821374e-01
5.53738713e-01 8.37388873e-01 -1.45844564e-01 -3.16188037e-01
8.85269344e-01 4.98569369e-01 1.22601402e+00 -1.04457724e+00
-1.04569328e+00 2.90368587e-01 2.07547229e-02 -8.78370047e-01
-2.27487311e-01 -9.27295625e-01 -1.08105922e+00 4.21991825e-01
-3.35208923e-01 -4.38631773e-01 1.00766242e+00 1.14517713e+00
3.15771490e-01 5.20680308e-01 -1.52502328e-01 -1.63843501e+00
-7.15067148e-01 -1.46501637e+00 -6.53221607e-01 3.06451321e-01
1.63968369e-01 -9.79988635e-01 3.92034128e-02 3.07250410e-01]
|
[6.548125267028809, -0.7172561287879944]
|
cfa8a915-e680-4a60-b6a3-3fcb916ef910
|
medical-image-segmentation-using-squeeze-and
|
2105.09511
| null |
https://arxiv.org/abs/2105.09511v3
|
https://arxiv.org/pdf/2105.09511v3.pdf
|
Medical Image Segmentation Using Squeeze-and-Expansion Transformers
|
Medical image segmentation is important for computer-aided diagnosis. Good segmentation demands the model to see the big picture and fine details simultaneously, i.e., to learn image features that incorporate large context while keep high spatial resolutions. To approach this goal, the most widely used methods -- U-Net and variants, extract and fuse multi-scale features. However, the fused features still have small "effective receptive fields" with a focus on local image cues, limiting their performance. In this work, we propose Segtran, an alternative segmentation framework based on transformers, which have unlimited "effective receptive fields" even at high feature resolutions. The core of Segtran is a novel Squeeze-and-Expansion transformer: a squeezed attention block regularizes the self attention of transformers, and an expansion block learns diversified representations. Additionally, we propose a new positional encoding scheme for transformers, imposing a continuity inductive bias for images. Experiments were performed on 2D and 3D medical image segmentation tasks: optic disc/cup segmentation in fundus images (REFUGE'20 challenge), polyp segmentation in colonoscopy images, and brain tumor segmentation in MRI scans (BraTS'19 challenge). Compared with representative existing methods, Segtran consistently achieved the highest segmentation accuracy, and exhibited good cross-domain generalization capabilities. The source code of Segtran is released at https://github.com/askerlee/segtran.
|
['Rick Goh', 'Yong liu', 'Xinxing Xu', 'Xiangde Luo', 'Xiuchao Sui', 'Shaohua Li']
|
2021-05-20
| null | null | null | null |
['optic-cup-segmentation']
|
['medical']
|
[ 2.91905910e-01 2.83425450e-01 -3.34457666e-01 -3.83733451e-01
-9.04362500e-01 -3.05478722e-01 7.60654211e-02 8.96228030e-02
-3.17198217e-01 4.83707964e-01 1.95672169e-01 -3.32805127e-01
-2.04332843e-01 -6.85464382e-01 -6.30439460e-01 -8.72464180e-01
-1.02294860e-02 1.57030925e-01 3.95808160e-01 -9.91721600e-02
1.33588061e-01 2.04627752e-01 -1.12958622e+00 5.44131100e-01
1.35915554e+00 1.11397111e+00 5.71941972e-01 4.67832357e-01
-4.53656130e-02 6.16182864e-01 -2.03088045e-01 -8.67504030e-02
2.52615929e-01 -4.18575704e-01 -9.36487734e-01 2.17398718e-01
3.70711148e-01 -2.49850795e-01 -2.58829623e-01 1.29472959e+00
6.08150125e-01 -4.74166982e-02 5.28689802e-01 -5.77744782e-01
-8.76237988e-01 5.29010296e-01 -7.44023025e-01 6.21667922e-01
6.45140111e-02 1.95371196e-01 7.11160183e-01 -4.61518139e-01
4.70588058e-01 1.03919947e+00 6.02083027e-01 5.90911865e-01
-1.02653706e+00 -5.26796043e-01 3.49824697e-01 -1.71102360e-01
-1.16267383e+00 -1.48556694e-01 4.69269216e-01 -4.83394921e-01
8.51088941e-01 3.99253339e-01 7.82695472e-01 8.02144766e-01
3.55427116e-01 1.07160676e+00 1.07599008e+00 -2.26058528e-01
1.64201409e-02 1.46156289e-02 3.14309269e-01 7.87532747e-01
2.18684345e-01 -6.91867201e-03 -5.05770445e-02 5.88540137e-02
1.21436393e+00 3.50782126e-01 -5.69719255e-01 -2.81845361e-01
-1.15870476e+00 7.55684972e-01 1.08600557e+00 6.00239277e-01
-3.81755173e-01 -2.19631985e-01 1.46656081e-01 5.50299473e-02
4.39049512e-01 5.41510165e-01 -3.58668149e-01 3.03348780e-01
-7.74737954e-01 -4.16106544e-02 3.01113427e-01 7.99504817e-01
5.82474530e-01 -3.51651847e-01 -3.60004812e-01 8.07451546e-01
2.12415144e-01 1.95262238e-01 1.08609700e+00 -6.19678497e-01
2.78259009e-01 9.18531656e-01 -2.29955822e-01 -7.22468495e-01
-6.42643213e-01 -8.03727031e-01 -1.05339515e+00 5.97973466e-02
3.63558084e-01 -3.90965603e-02 -1.47467899e+00 1.57427764e+00
3.10080439e-01 1.78627029e-01 1.91693082e-02 1.08331895e+00
1.08465278e+00 4.10838157e-01 -3.11445408e-02 2.32764124e-03
1.54562223e+00 -1.15838528e+00 -5.41966021e-01 -4.94080454e-01
6.72733843e-01 -5.12827277e-01 1.06159210e+00 2.77092457e-01
-1.24942172e+00 -5.08898199e-01 -1.06805265e+00 -3.78590167e-01
-4.27628458e-01 1.72300622e-01 7.56711841e-01 3.84350985e-01
-1.17765117e+00 4.59143728e-01 -1.07688260e+00 -1.95982158e-01
8.55833352e-01 6.60108089e-01 -2.78761715e-01 -1.77908599e-01
-1.02840722e+00 5.34082770e-01 3.16958338e-01 1.13104418e-01
-5.19531727e-01 -6.95051968e-01 -9.72012818e-01 -7.15752244e-02
3.05678248e-01 -9.25918221e-01 1.25330782e+00 -1.15459359e+00
-1.31069267e+00 9.91448283e-01 -6.63741380e-02 -6.04129493e-01
4.01704013e-01 -7.80672058e-02 -1.73256218e-01 4.03160661e-01
2.93894112e-01 1.11107111e+00 6.31633461e-01 -1.01704431e+00
-7.01836586e-01 -6.56027496e-01 1.61461085e-01 3.89624655e-01
-1.72964916e-01 -3.06146801e-01 -6.05215371e-01 -7.00111032e-01
3.76093060e-01 -7.15792477e-01 -6.13197803e-01 -2.15889379e-01
-4.66621280e-01 -7.23952129e-02 4.68164384e-01 -6.71306312e-01
1.17583191e+00 -2.31480193e+00 6.66397810e-02 3.98298167e-02
6.49537861e-01 2.82805920e-01 -3.51785235e-02 -3.36585909e-01
-1.05943859e-01 2.28190824e-01 -4.13036913e-01 -1.72192678e-02
-3.73268574e-01 6.75708577e-02 1.72433227e-01 3.37984830e-01
1.60229191e-01 1.14903164e+00 -8.27298045e-01 -5.08912861e-01
2.22220391e-01 3.70878279e-01 -7.93114126e-01 -5.20776063e-02
-2.39445381e-02 6.90768242e-01 -8.17581952e-01 8.66957903e-01
8.06600094e-01 -7.84245074e-01 -5.32766357e-02 -3.81263822e-01
-1.43111631e-01 5.68772964e-02 -7.28582323e-01 2.00439191e+00
-2.43655160e-01 3.16844881e-01 1.43303692e-01 -1.08189809e+00
6.46600485e-01 2.37785876e-01 7.00256467e-01 -9.62438047e-01
4.06486452e-01 3.84672910e-01 1.89193115e-01 -6.40920222e-01
8.78322050e-02 -1.12303860e-01 -1.13238998e-01 -1.50156274e-01
4.51377518e-02 9.65618156e-03 1.30407840e-01 -4.53383923e-02
8.37550461e-01 -1.65542394e-01 4.78185892e-01 -3.93581182e-01
4.26812440e-01 1.26643730e-02 6.73309803e-01 6.24509811e-01
-2.82017887e-01 8.96717906e-01 5.81705630e-01 -4.22493368e-01
-5.85747540e-01 -8.83079529e-01 -4.26070660e-01 6.46758735e-01
5.71589649e-01 -1.01035580e-01 -8.85598719e-01 -8.08281958e-01
-6.48658648e-02 1.51111722e-01 -9.13644493e-01 -9.68813673e-02
-6.12160981e-01 -9.07905936e-01 2.82196552e-01 6.85118675e-01
7.60086656e-01 -9.27620828e-01 -9.32638466e-01 1.68775067e-01
-2.74998546e-01 -9.52263057e-01 -7.79541731e-01 2.42341816e-01
-1.06562209e+00 -1.14938653e+00 -1.10578442e+00 -1.18693566e+00
9.26097035e-01 2.70549983e-01 1.02340198e+00 -1.31800309e-01
-6.65979445e-01 4.04802375e-02 -3.00832331e-01 -3.33828956e-01
5.06953262e-02 1.53955176e-01 -4.29134965e-01 -1.55320063e-01
2.93884248e-01 -4.51761693e-01 -1.02791047e+00 3.68607342e-01
-1.01229441e+00 3.73051614e-01 1.12933505e+00 1.16439652e+00
9.51055765e-01 -1.38929442e-01 4.59238738e-01 -9.68870997e-01
4.50494379e-01 -4.24840748e-01 -4.45634782e-01 1.94079265e-01
-3.41483682e-01 -1.87505394e-01 4.21398997e-01 -2.46283129e-01
-9.86240923e-01 -2.85009108e-03 -3.52609932e-01 -4.37344015e-01
-1.53561607e-01 5.04531562e-01 -1.03967659e-01 -1.39837697e-01
6.70699835e-01 2.73488969e-01 1.27789602e-01 -4.32998955e-01
9.50347036e-02 5.43172896e-01 5.53118825e-01 -2.85858363e-01
1.13085406e-02 5.74719012e-01 -3.94464165e-01 -6.39657497e-01
-8.17231417e-01 -5.97012281e-01 -5.62301278e-01 3.21207732e-01
1.04700530e+00 -8.89764369e-01 -3.94593149e-01 5.20773351e-01
-6.99472308e-01 -4.29373324e-01 -4.93185103e-01 6.66469276e-01
-4.34638858e-01 1.98754609e-01 -6.80866659e-01 -2.02478856e-01
-4.04629588e-01 -1.67199099e+00 1.07888043e+00 6.96674407e-01
2.04600260e-01 -9.57933903e-01 -2.48365775e-01 4.23825085e-01
4.18631345e-01 3.43508691e-01 8.42991650e-01 -6.35959327e-01
-7.71870017e-01 8.54067951e-02 -5.51217616e-01 4.26073283e-01
3.09783787e-01 -5.23757041e-01 -8.37506890e-01 -4.01505470e-01
9.80600268e-02 -1.87640831e-01 1.22791588e+00 1.22224295e+00
1.48514783e+00 -1.05251297e-01 -6.62901402e-01 1.15857291e+00
1.36483014e+00 5.13146102e-01 5.50081789e-01 1.59460381e-01
5.96996188e-01 3.71367484e-01 3.63211095e-01 2.04692557e-01
3.83528709e-01 3.11806560e-01 4.04210180e-01 -6.73604488e-01
-2.78190225e-01 -2.44322009e-02 -1.44056335e-01 6.06480181e-01
-1.40104825e-02 -1.44797847e-01 -8.93797040e-01 7.29148328e-01
-1.68800747e+00 -4.11976725e-01 2.74115473e-01 1.86551571e+00
7.49963403e-01 4.81147766e-02 -4.35381718e-02 -2.30327055e-01
5.98760962e-01 -2.02123472e-03 -8.73721123e-01 -1.23023884e-02
-2.74598282e-02 2.98675179e-01 5.50103664e-01 3.72866690e-01
-1.36678910e+00 7.40679920e-01 5.50787306e+00 9.37814236e-01
-1.42725265e+00 2.18752638e-01 1.06598377e+00 1.42156053e-02
-2.30325580e-01 -3.43083322e-01 -7.09681094e-01 4.38559443e-01
3.58392358e-01 1.24626495e-01 -7.99832493e-02 5.71627975e-01
-8.13222677e-02 -2.00165674e-01 -8.54840457e-01 9.44745302e-01
-2.42143217e-02 -1.50607288e+00 -2.84492839e-02 2.47818768e-01
7.29179978e-01 2.66049415e-01 4.13887382e-01 1.35633945e-01
9.69155431e-02 -1.18389547e+00 2.27370203e-01 4.68077749e-01
9.81495917e-01 -3.57999414e-01 7.98527360e-01 1.72533035e-01
-1.12359440e+00 -1.87408313e-01 -3.18704754e-01 3.61468762e-01
-1.06434092e-01 4.92846668e-01 -7.39476860e-01 5.31293333e-01
7.59518385e-01 9.60567355e-01 -5.54643989e-01 1.44443679e+00
9.45329666e-02 3.15791398e-01 -2.15062350e-01 1.84987158e-01
6.45418286e-01 -1.21152110e-01 4.06407475e-01 1.19054222e+00
4.53661770e-01 3.11538279e-01 3.71625632e-01 8.04021537e-01
-5.52972555e-02 1.98799640e-01 -3.77668053e-01 2.18245581e-01
1.03407346e-01 1.16942048e+00 -9.75572944e-01 -3.86915207e-01
-4.80388463e-01 8.31325293e-01 -4.86494508e-03 4.12248552e-01
-7.51068413e-01 -3.32487166e-01 3.69325131e-01 1.51177257e-01
3.12449217e-01 3.47308874e-01 -4.31367785e-01 -1.26872134e+00
-9.22644958e-02 -9.34038639e-01 5.98788500e-01 -4.85228151e-01
-1.11170268e+00 8.90978634e-01 -9.57966074e-02 -1.32309651e+00
1.12775065e-01 -7.28668153e-01 -3.95175517e-01 8.05130720e-01
-1.74360955e+00 -1.28174472e+00 -5.05636156e-01 7.15372384e-01
6.53848588e-01 3.99244903e-03 5.39232373e-01 3.82710725e-01
-6.04333937e-01 6.33788049e-01 1.06485456e-01 1.84547603e-01
5.34412563e-01 -1.33538222e+00 1.54324412e-01 5.41480720e-01
-2.01758832e-01 5.58756948e-01 1.06006883e-01 -5.58907568e-01
-8.72580588e-01 -1.12406421e+00 3.39753419e-01 -6.50701299e-02
3.13201338e-01 2.94163805e-02 -1.05037522e+00 7.09028721e-01
2.77113378e-01 3.45320165e-01 6.57872200e-01 -2.31827527e-01
-7.05782548e-02 -4.26891111e-02 -1.27085555e+00 3.62590909e-01
9.66686308e-01 -4.12679203e-02 -6.07642531e-01 3.63418728e-01
7.68699169e-01 -9.17410314e-01 -9.14648831e-01 6.38852239e-01
4.68155175e-01 -1.03732252e+00 1.06089187e+00 -2.56442398e-01
3.88987005e-01 -1.31290317e-01 3.05668935e-02 -1.20114851e+00
-5.05475223e-01 -4.57205892e-01 3.43948275e-01 6.35981023e-01
4.23630357e-01 -1.00441349e+00 6.85214579e-01 2.39832222e-01
-6.31648004e-01 -1.19761968e+00 -8.63173962e-01 -3.35195243e-01
2.68047094e-01 -6.16174471e-03 6.11616373e-01 8.90060663e-01
4.99228649e-02 1.56695545e-02 1.26880392e-01 1.26215532e-01
4.01402861e-01 2.29565978e-01 6.75589591e-02 -9.45842564e-01
-3.29074830e-01 -8.23809206e-01 -3.90508443e-01 -1.32726192e+00
-3.19861352e-01 -1.05880344e+00 -1.72109246e-01 -1.75847375e+00
3.69539052e-01 -5.88551939e-01 -6.01427674e-01 5.75942934e-01
-2.39426464e-01 2.36861810e-01 -7.10675940e-02 2.49976426e-01
-4.89586711e-01 2.45180771e-01 1.87134254e+00 -2.58513212e-01
-3.60302806e-01 2.45314866e-01 -1.12168622e+00 8.65244567e-01
8.79831016e-01 -5.02259545e-02 -5.80836952e-01 -5.18237591e-01
-2.80962378e-01 2.16267571e-01 3.86115998e-01 -9.87155795e-01
3.13498259e-01 9.32451785e-02 5.89752257e-01 -5.99046648e-01
1.36402726e-01 -4.49096352e-01 -2.26368979e-01 5.76055527e-01
-3.33880544e-01 -2.08150029e-01 3.33574682e-01 3.25174630e-01
-6.25110805e-01 4.75369692e-02 9.99233246e-01 -5.42405903e-01
-7.17105806e-01 6.10896170e-01 -1.66203752e-01 1.54345125e-01
1.04769397e+00 -3.48091424e-01 -3.49692643e-01 1.78587988e-01
-9.88559425e-01 5.11521935e-01 3.09847742e-01 2.10634291e-01
8.20363641e-01 -9.88235950e-01 -5.91152966e-01 5.55008292e-01
-1.56055177e-02 6.06350362e-01 6.99007094e-01 1.37932861e+00
-4.92589504e-01 5.86852312e-01 -1.93872303e-01 -8.45173061e-01
-9.53078210e-01 4.58845019e-01 7.81077385e-01 -5.33790290e-01
-9.03988421e-01 1.24612510e+00 1.05387831e+00 -2.78651267e-01
5.28029092e-02 -1.02740216e+00 -4.42727357e-01 3.52011509e-02
5.24730325e-01 -1.78111166e-01 1.05921201e-01 -4.79945272e-01
-2.69611329e-01 7.62447953e-01 -4.08004612e-01 3.52655023e-01
1.13521302e+00 -1.89867795e-01 -2.05298830e-02 -3.80337760e-02
1.15484381e+00 -1.85624301e-01 -1.27746344e+00 -4.02982533e-01
-2.82199353e-01 -5.05443990e-01 3.81759167e-01 -1.01886964e+00
-1.31150615e+00 9.77410078e-01 9.56265032e-01 2.94433832e-01
1.50369751e+00 1.92836300e-01 1.00557733e+00 -1.50164828e-01
1.04451753e-01 -6.60077870e-01 -4.20183055e-02 1.75711706e-01
7.68638849e-01 -1.37878227e+00 -2.66297132e-01 -5.07584870e-01
-7.60349512e-01 8.80283833e-01 8.02512705e-01 -1.02980718e-01
6.63067997e-01 2.89110065e-01 -1.76053867e-02 -2.59838045e-01
-4.78230268e-01 -3.69150460e-01 5.81856430e-01 4.89008725e-01
5.04962623e-01 2.34322280e-01 -1.34617522e-01 9.50538218e-01
6.44756779e-02 2.28475720e-01 2.16410115e-01 7.37618029e-01
-5.52692473e-01 -8.24404061e-01 -1.96383759e-01 7.57370591e-01
-6.33301139e-01 -2.17965007e-01 4.21970189e-02 8.10878396e-01
5.05719244e-01 5.90385914e-01 1.27492473e-01 -1.67669296e-01
3.27355474e-01 -3.76546562e-01 4.07331765e-01 -6.70347214e-01
-5.80552220e-01 6.23636663e-01 -2.62487203e-01 -5.84509790e-01
-3.46655011e-01 -5.36734581e-01 -1.42716956e+00 3.20977718e-01
-2.20153511e-01 1.76511213e-01 2.83749044e-01 6.90077186e-01
4.01009858e-01 9.81513798e-01 4.57070172e-01 -6.77946627e-01
-3.08070391e-01 -8.46081734e-01 -6.57548130e-01 2.27536246e-01
5.69980860e-01 -5.44838428e-01 -1.68958172e-01 -2.39229668e-02]
|
[14.665172576904297, -2.5833194255828857]
|
33cc0e1c-4cf9-4f60-a2ce-d805544f92dc
|
neuspike-net-high-speed-video-reconstruction
| null | null |
http://openaccess.thecvf.com//content/ICCV2021/html/Zhu_NeuSpike-Net_High_Speed_Video_Reconstruction_via_Bio-Inspired_Neuromorphic_Cameras_ICCV_2021_paper.html
|
http://openaccess.thecvf.com//content/ICCV2021/papers/Zhu_NeuSpike-Net_High_Speed_Video_Reconstruction_via_Bio-Inspired_Neuromorphic_Cameras_ICCV_2021_paper.pdf
|
NeuSpike-Net: High Speed Video Reconstruction via Bio-Inspired Neuromorphic Cameras
|
Neuromorphic vision sensor is a new bio-inspired imaging paradigm that emerged in recent years, which continuously sensing luminance intensity and firing asynchronous spikes (events) with high temporal resolution. Typically, there are two types of neuromorphic vision sensors, namely dynamic vision sensor (DVS) and spike camera. From the perspective of bio-inspired sampling, DVS only perceives movement by imitating the retinal periphery, while the spike camera was developed to perceive fine textures by simulating the fovea. It is meaningful to explore how to combine two types of neuromorphic cameras to reconstruct high quality image like human vision. In this paper, we propose a NeuSpike-Net to learn both the high dynamic range and high motion sensitivity of DVS and the full texture sampling of spike camera to achieve high-speed and high dynamic image reconstruction. We propose a novel representation to effectively extract the temporal information of spike and event data. By introducing the feature fusion module, the two types of neuromorphic data achieve complementary to each other. The experimental results on the simulated and real datasets demonstrate that the proposed approach is effective to reconstruct high-speed and high dynamic range images via the combination of spike and event data.
|
['Yonghong Tian', 'Tiejun Huang', 'Xiao Wang', 'Jianing Li', 'Lin Zhu']
|
2021-01-01
| null | null | null |
iccv-2021-1
|
['video-reconstruction']
|
['computer-vision']
|
[ 5.31980574e-01 -7.99946189e-01 6.38638139e-01 -1.93281248e-01
-6.68905228e-02 -4.71552312e-01 5.93027651e-01 -5.31178832e-01
-6.84894919e-01 6.41335845e-01 -1.82912499e-01 3.79874825e-01
-3.87453772e-02 -6.18792176e-01 -7.92084038e-01 -1.18051219e+00
4.15651590e-01 -3.95928293e-01 8.29005122e-01 1.22363754e-01
5.35266936e-01 3.30299228e-01 -2.18383121e+00 5.23279488e-01
6.00853682e-01 1.48475051e+00 9.57158566e-01 5.67754030e-01
-6.68425187e-02 5.40840566e-01 -5.52624822e-01 3.50311249e-01
2.96557307e-01 -6.14713550e-01 3.47942144e-01 2.75972597e-02
-5.49563803e-02 -2.38444299e-01 -5.55806696e-01 1.24009216e+00
5.90899348e-01 -4.37229067e-01 3.35241556e-01 -1.00671804e+00
-7.94618189e-01 1.27787724e-01 -4.90592241e-01 5.60957134e-01
2.73177534e-01 7.00943887e-01 3.01588364e-02 -5.90559065e-01
6.06363833e-01 8.73607934e-01 4.22210515e-01 6.95430458e-01
-9.63239670e-01 -5.93111575e-01 -2.23940730e-01 4.83588755e-01
-1.10757411e+00 -4.92127985e-01 7.01074958e-01 -2.10680336e-01
9.15507913e-01 8.08535293e-02 1.15065491e+00 1.29241049e+00
8.20483148e-01 4.58837777e-01 1.88021123e+00 7.83263296e-02
5.80916047e-01 -3.48967761e-01 2.54531145e-01 4.09341216e-01
3.71246070e-01 3.37910682e-01 -7.67908514e-01 5.86986765e-02
1.14790857e+00 4.52930629e-01 -7.07708895e-01 1.07894033e-01
-1.30203903e+00 -1.06134312e-02 3.08272988e-01 2.67761678e-01
-4.87197995e-01 3.39897454e-01 -5.64169604e-03 1.83953211e-01
-5.17499685e-01 2.32045185e-02 -3.77157591e-02 -2.50844389e-01
-6.14456117e-01 -9.55945253e-02 6.97623253e-01 8.02695155e-01
6.22836530e-01 2.85973847e-01 -2.10191578e-01 4.16541189e-01
2.71099716e-01 9.38318968e-01 1.13322365e+00 -1.07450116e+00
-2.56039500e-01 7.31690884e-01 5.74914878e-03 -6.17831707e-01
-2.00230807e-01 -5.20589985e-02 -1.02093923e+00 5.32737017e-01
2.47856826e-01 1.66542992e-01 -1.24840748e+00 1.58898461e+00
-2.57874072e-01 6.96318328e-01 1.61675364e-01 1.21111214e+00
7.15407729e-01 7.69455552e-01 -2.91820168e-01 -5.59506178e-01
1.44401753e+00 -2.92792052e-01 -8.36480916e-01 -2.66604662e-01
-4.58018273e-01 -3.04815024e-01 8.43185067e-01 3.11739445e-01
-1.16521096e+00 -6.37479663e-01 -1.28085804e+00 1.07525319e-01
-3.77084762e-01 -1.09484173e-01 4.27585244e-01 1.80056840e-01
-1.06329215e+00 3.62626493e-01 -9.59142864e-01 -3.46209675e-01
3.39232326e-01 4.38007563e-01 -1.96443915e-01 1.04931081e-02
-7.69645333e-01 5.66883087e-01 2.53063351e-01 -4.20233756e-02
-9.69139874e-01 -2.82801211e-01 -1.62227780e-01 4.96360622e-02
-2.51489162e-01 -9.89476740e-01 8.87555778e-01 -9.83061731e-01
-1.52502680e+00 8.94650996e-01 -3.14918965e-01 -5.64542770e-01
3.93652990e-02 6.23845458e-01 -3.76401186e-01 3.88864607e-01
-2.48405650e-01 3.85570139e-01 7.89263308e-01 -9.79879320e-01
-5.41978240e-01 -8.78582299e-01 -4.30436492e-01 -4.36712243e-03
-2.68770099e-01 -8.48658904e-02 -3.34685713e-01 -2.15033323e-01
2.49105513e-01 -5.48384607e-01 1.50939852e-01 1.98482066e-01
2.39067227e-01 1.19153798e-01 8.70714366e-01 -2.30211645e-01
7.28173733e-01 -2.30350828e+00 3.99582759e-02 -2.77406603e-01
2.19344482e-01 4.45676804e-01 -1.93862319e-01 1.72860220e-01
3.27816844e-01 -5.68172932e-01 -5.16182363e-01 1.04961872e-01
-5.23094177e-01 3.08977187e-01 -4.16027874e-01 3.83834660e-01
1.15649574e-01 1.05722570e+00 -6.76775694e-01 -2.29010731e-01
2.19379067e-01 5.58747590e-01 8.28650221e-02 1.49071500e-01
-1.29114881e-01 6.87488496e-01 -3.55019659e-01 8.64945292e-01
8.22993815e-01 -1.80953696e-01 -1.00370571e-01 -2.71218628e-01
-5.07135868e-01 -2.44346172e-01 -9.26135004e-01 1.79883289e+00
2.08333191e-02 6.36214793e-01 7.39827082e-02 -6.80445910e-01
1.21495020e+00 1.35644525e-01 3.28239083e-01 -1.35525191e+00
3.90616268e-01 4.04580563e-01 6.64301366e-02 -8.03459823e-01
-2.48441264e-01 -7.30187669e-02 3.02956074e-01 9.17227790e-02
6.64422810e-02 8.30941461e-03 -1.11759022e-01 -3.69796872e-01
1.41919506e+00 3.89276743e-02 -3.62472199e-02 -1.10034227e-01
4.09467459e-01 -2.61143357e-01 7.79850304e-01 5.49227655e-01
-2.87224412e-01 7.08591461e-01 -8.17188621e-02 -3.18431526e-01
-9.54878151e-01 -1.54595017e+00 -1.93980545e-01 1.53499126e-01
7.70327330e-01 3.42632800e-01 -7.62469828e-01 4.70050901e-01
-1.31364658e-01 5.80287203e-02 -3.67744088e-01 -3.08442175e-01
-4.36435223e-01 -6.53414011e-01 4.92731899e-01 3.90098155e-01
1.08689129e+00 -1.28359520e+00 -1.33717275e+00 3.29905182e-01
1.48503900e-01 -1.23361051e+00 -1.04681917e-01 3.17684352e-01
-8.54331851e-01 -1.00263572e+00 -7.37689912e-01 -1.10752499e+00
4.43805397e-01 4.02080715e-01 3.58779371e-01 -4.71759379e-01
-8.43376637e-01 4.00282145e-01 2.56318715e-03 -4.58388418e-01
2.74317533e-01 -7.03188360e-01 5.77377193e-02 4.48582530e-01
5.49338937e-01 -1.21048868e+00 -9.02214408e-01 4.72038053e-02
-1.11994052e+00 1.50398746e-01 7.49949038e-01 6.20890617e-01
1.00911176e+00 -1.37675330e-01 3.28247041e-01 -3.41048867e-01
5.97735941e-01 -1.91652641e-01 -7.42541850e-01 1.62483007e-01
-3.75272423e-01 -9.33527127e-02 8.76828551e-01 -8.15859973e-01
-9.07692015e-01 2.18907565e-01 2.96009988e-01 -5.91325641e-01
-2.81573564e-01 1.36732280e-01 -7.19340816e-02 -3.59267265e-01
5.70273221e-01 1.17832732e+00 1.03007488e-01 -3.21281701e-01
-3.25383544e-01 9.04595256e-01 1.17365980e+00 -1.01693263e-02
3.52516383e-01 9.59703922e-01 1.12613784e-02 -7.26557016e-01
2.70592477e-02 -2.68823773e-01 -1.57474130e-02 -4.03878659e-01
8.13248992e-01 -9.23609912e-01 -1.23880720e+00 1.35439622e+00
-1.28137255e+00 -8.38327035e-02 -2.85756737e-01 5.77903509e-01
-8.30720425e-01 1.85230941e-01 -7.40932107e-01 -7.92862475e-01
-3.26067865e-01 -9.63394880e-01 9.36117351e-01 9.19084132e-01
6.02111459e-01 -3.13295215e-01 2.05881134e-01 -1.98645741e-01
4.56901491e-01 4.25903946e-01 5.29876649e-01 2.43957669e-01
-1.11211932e+00 1.03307073e-03 -4.34038997e-01 2.39994917e-02
5.00265919e-02 -1.35844246e-01 -1.24422264e+00 -1.03329122e-01
7.52702355e-01 -6.29857183e-02 1.16560161e+00 4.65712756e-01
1.19325769e+00 6.90566078e-02 -2.65698463e-01 1.08300042e+00
1.89836884e+00 7.00489402e-01 1.04944098e+00 7.57165551e-02
2.52643526e-01 2.04787716e-01 2.31473800e-02 5.02158761e-01
2.57653266e-01 3.34652007e-01 5.07144749e-01 3.51949453e-01
-2.96127945e-01 -2.21246690e-01 4.86394078e-01 1.05809569e+00
-2.97411680e-01 -2.57820431e-02 -5.71211696e-01 4.79208171e-01
-1.94353604e+00 -1.09771800e+00 -1.31303027e-01 2.21803236e+00
7.94818938e-01 1.09370034e-02 -2.08991870e-01 1.98449250e-02
9.17668879e-01 -2.27958679e-01 -1.17298973e+00 -6.24395758e-02
-7.61112750e-01 2.91542500e-01 5.46617329e-01 -1.43644959e-01
-3.20721298e-01 4.79576886e-01 6.07495308e+00 3.63828331e-01
-1.44281363e+00 -6.50034547e-02 -3.65984365e-02 -2.98722267e-01
-2.90046901e-01 2.16503628e-02 -6.94921196e-01 1.21249592e+00
9.72857296e-01 -2.31514961e-01 9.66370285e-01 1.25963211e-01
1.97052315e-01 -3.62553358e-01 -7.83010840e-01 1.62558508e+00
1.16638448e-02 -1.37487793e+00 1.34313568e-01 1.25209568e-02
5.24179459e-01 6.06541671e-02 1.94022000e-01 -3.45536023e-01
-2.04778776e-01 -7.63271451e-01 5.02208650e-01 1.42560935e+00
7.92183578e-01 -2.53479183e-01 3.79676223e-01 6.31508291e-01
-1.09877658e+00 -5.46178341e-01 -6.80609822e-01 -3.78432363e-01
2.01066788e-02 6.05358481e-01 1.67628913e-03 1.66756343e-02
1.03744531e+00 9.28539991e-01 -4.21870768e-01 1.52148283e+00
2.96138734e-01 2.81045258e-01 -3.24723691e-01 -3.75509560e-01
-1.26558289e-01 -3.03268552e-01 5.99822760e-01 8.08307350e-01
7.06748188e-01 5.11532128e-01 -2.96922296e-01 1.37854362e+00
1.71228066e-01 -6.52710974e-01 -8.56710255e-01 -1.63363665e-01
6.35036290e-01 1.21616638e+00 -6.06943548e-01 -1.24576904e-01
-3.61807585e-01 1.18880808e+00 -3.69078703e-02 4.10986394e-01
-7.16547251e-01 -6.08277559e-01 4.60422546e-01 -7.44299730e-03
3.86908323e-01 -3.98762852e-01 -4.10564035e-01 -1.22232115e+00
2.47581750e-01 -3.39769423e-01 -2.36765459e-01 -1.32584834e+00
-9.75086927e-01 4.54481483e-01 -5.60130775e-01 -1.33267093e+00
9.16780755e-02 -8.76041174e-01 -6.59101367e-01 8.03712368e-01
-1.54582489e+00 -8.00707519e-01 -7.69856989e-01 1.06131399e+00
2.80289769e-01 -9.47737470e-02 6.61577642e-01 -1.16696600e-02
-3.01916838e-01 9.24470276e-02 3.99511099e-01 -1.68333232e-01
5.17960489e-01 -8.23447585e-01 7.56010935e-02 8.02208126e-01
-6.46246150e-02 4.66013342e-01 2.73288548e-01 -4.44748044e-01
-2.02828979e+00 -7.54192412e-01 2.36517072e-01 1.18703380e-01
3.65735114e-01 -3.84272665e-01 -8.81581068e-01 1.54610410e-01
1.45587549e-01 7.65142739e-02 2.12650165e-01 -1.14886272e+00
-4.53079373e-01 -7.10032642e-01 -1.31909227e+00 5.44215739e-01
1.24102747e+00 -6.01786315e-01 -7.67787397e-01 -1.38172582e-01
4.46106583e-01 -4.16312292e-02 -5.54837108e-01 4.05713528e-01
7.41970360e-01 -1.34872377e+00 5.96853554e-01 2.66015291e-01
1.92001805e-01 -7.09302306e-01 -2.18562335e-01 -8.86240065e-01
-2.60848969e-01 -3.55860710e-01 -2.08608970e-01 1.01705825e+00
-2.81798542e-01 -1.09491217e+00 3.74688953e-01 7.22423717e-02
-1.02540359e-01 -5.22471011e-01 -1.10677934e+00 -8.58483493e-01
-5.71647584e-01 1.06929772e-01 3.43517035e-01 2.75059879e-01
-1.46178588e-01 2.55349819e-02 9.17190388e-02 1.65227979e-01
7.98310280e-01 4.03519303e-01 7.52318799e-02 -1.20411003e+00
-4.18079913e-01 -3.27519864e-01 -1.01967871e+00 -1.01393521e+00
-2.72012204e-01 -7.03924060e-01 2.58343905e-01 -1.48842049e+00
4.85272855e-01 3.69837791e-01 -6.77448273e-01 7.19772652e-02
3.17719966e-01 4.39149290e-01 6.96159080e-02 4.98475283e-01
-3.79237533e-01 3.49975526e-01 1.29423642e+00 -1.83781702e-02
-3.54175689e-03 -3.51651937e-01 -3.36463451e-01 5.69832087e-01
5.09059727e-01 -1.03445649e-01 -3.43861908e-01 -6.71627402e-01
-6.38907999e-02 3.09363604e-01 8.43664050e-01 -1.63163030e+00
1.05971134e+00 -2.57734656e-02 7.62847900e-01 -2.97264159e-01
4.45017844e-01 -7.41488457e-01 4.50484514e-01 6.28077686e-01
-3.63153368e-02 2.52953153e-02 1.08208410e-01 9.24827933e-01
-3.21179897e-01 2.51736175e-02 1.07688284e+00 -6.29808486e-01
-9.78900969e-01 2.77466923e-01 -6.76811814e-01 -1.62734956e-01
1.10302508e+00 -9.08388197e-01 -7.55786777e-01 8.51588883e-03
-3.75199467e-01 -1.28880307e-01 7.61924565e-01 1.38509616e-01
1.25446284e+00 -1.24725461e+00 -2.23605484e-01 6.97719514e-01
9.48011503e-02 -4.53123987e-01 3.59340847e-01 6.05790675e-01
-3.77819180e-01 1.46652982e-01 -1.08165920e+00 -8.33941996e-01
-7.98681855e-01 5.68380117e-01 4.89212602e-01 5.02446949e-01
-6.70901895e-01 5.73877335e-01 1.20433800e-01 2.12451443e-01
1.64887786e-01 -2.79651552e-01 -5.67259602e-02 -3.75459880e-01
7.64227211e-01 2.54419714e-01 -3.50023329e-01 -1.73546776e-01
-4.04072642e-01 1.13202727e+00 3.10165942e-01 -3.72520745e-01
1.33107662e+00 -3.07495117e-01 -3.71092588e-01 9.55498457e-01
1.04110920e+00 -5.41760862e-01 -1.59059393e+00 -2.83655636e-02
-2.29866862e-01 -3.16076696e-01 -1.82265580e-01 -9.70904410e-01
-1.00456655e+00 1.11112273e+00 1.08946359e+00 1.71162456e-01
1.82993412e+00 -4.00667578e-01 1.06780982e+00 3.60214710e-01
8.16722453e-01 -1.03570628e+00 1.99000880e-01 3.21455032e-01
6.24759555e-01 -6.85947835e-01 -6.29635930e-01 -3.20661739e-02
-2.54022390e-01 1.29537880e+00 6.88231766e-01 -6.85731232e-01
5.48772991e-01 7.45081663e-01 -1.54011741e-01 -3.97797450e-02
-1.05219710e+00 -3.19218397e-01 -2.58284628e-01 8.51377726e-01
-2.09366903e-01 -1.19556494e-01 -3.77773583e-01 6.71600103e-01
2.55615264e-01 6.71163976e-01 6.52933121e-01 8.68410587e-01
-8.46522868e-01 -5.67269444e-01 -2.46314123e-01 4.58565205e-01
-1.71283498e-01 9.71546918e-02 -4.42269623e-01 1.62542045e-01
3.64275873e-01 6.70489013e-01 4.14489418e-01 -6.15594685e-01
2.74826497e-01 3.97391841e-02 8.89348984e-01 -1.65283173e-01
-2.92099088e-01 7.33822137e-02 -8.44991744e-01 -7.18936920e-01
-4.68066990e-01 -2.89613724e-01 -1.66725707e+00 2.44988054e-02
1.16028212e-01 -4.76894617e-01 9.74618912e-01 7.20783889e-01
6.39860272e-01 4.61769313e-01 6.95389390e-01 -5.82594812e-01
-3.95405233e-01 -6.59893930e-01 -9.37635601e-01 2.93051034e-01
3.68145198e-01 -3.54744077e-01 -5.18904030e-01 2.38155916e-01]
|
[8.274908065795898, 2.2212183475494385]
|
62c9cb1a-9398-4940-8744-916c7a81d130
|
monotdp-twin-depth-perception-for-monocular
|
2305.10974
| null |
https://arxiv.org/abs/2305.10974v2
|
https://arxiv.org/pdf/2305.10974v2.pdf
|
MonoTDP: Twin Depth Perception for Monocular 3D Object Detection in Adverse Scenes
|
3D object detection plays a crucial role in numerous intelligent vision systems. Detection in the open world inevitably encounters various adverse scenes, such as dense fog, heavy rain, and low light conditions. Although existing efforts primarily focus on diversifying network architecture or training schemes, resulting in significant progress in 3D object detection, most of these learnable modules fail in adverse scenes, thereby hindering detection performance. To address this issue, this paper proposes a monocular 3D detection model designed to perceive twin depth in adverse scenes, termed MonoTDP, which effectively mitigates the degradation of detection performance in various harsh environments. Specifically, we first introduce an adaptive learning strategy to aid the model in handling uncontrollable weather conditions, significantly resisting degradation caused by various degrading factors. Then, to address the depth/content loss in adverse regions, we propose a novel twin depth perception module that simultaneously estimates scene and object depth, enabling the integration of scene-level features and object-level features. Additionally, we assemble a new adverse 3D object detection dataset encompassing a wide range of challenging scenes, including rainy, foggy, and low light weather conditions, with each type of scene containing 7,481 images. Experimental results demonstrate that our proposed method outperforms current state-of-the-art approaches by an average of 3.12% in terms of AP_R40 for car category across various adverse environments.
|
['JinYuan Liu', 'Risheng Liu', 'Xin Fan', 'Long Ma', 'Yixin Lei', 'Xingyuan Li']
|
2023-05-18
| null | null | null | null |
['monocular-3d-object-detection']
|
['computer-vision']
|
[ 2.50610203e-01 -5.72477341e-01 3.42649817e-01 -3.47969979e-01
-1.74523637e-01 -4.86538947e-01 2.23251313e-01 -2.84978420e-01
-4.49538976e-01 3.28817248e-01 -2.60704964e-01 -1.55663908e-01
2.48007894e-01 -7.71934271e-01 -5.49889028e-01 -9.36748624e-01
3.12596336e-02 -1.31297246e-01 6.27225816e-01 -1.04129218e-01
2.16942638e-01 6.18083298e-01 -1.89150369e+00 -4.57415245e-02
1.05928838e+00 1.18233252e+00 8.57665181e-01 7.63589621e-01
5.19991815e-02 6.98871434e-01 -7.14713812e-01 -7.34008402e-02
6.54361665e-01 7.80584067e-02 1.00272976e-01 3.34273249e-01
9.60819960e-01 -1.13271844e+00 -4.53063101e-01 1.25300539e+00
7.05173314e-01 5.57222366e-02 4.77020383e-01 -1.07354879e+00
-6.94383919e-01 -3.75844300e-01 -9.16602552e-01 6.68927431e-01
3.32056820e-01 6.34881496e-01 4.83802080e-01 -1.04099524e+00
4.73393910e-02 1.44509077e+00 4.92699236e-01 6.11306012e-01
-5.85864186e-01 -7.52762973e-01 5.80616117e-01 2.98453927e-01
-1.09016812e+00 -3.03737581e-01 5.35745382e-01 -3.23990434e-01
1.03847158e+00 2.40512416e-02 5.98738909e-01 8.39109719e-01
3.27087581e-01 8.68345976e-01 1.14548361e+00 -1.47683442e-01
7.97561556e-02 8.43568519e-02 1.15846746e-01 6.24605954e-01
7.54782200e-01 3.88576239e-01 -2.35140845e-01 3.56965631e-01
6.42009854e-01 8.41620341e-02 -3.46295238e-01 -1.92510188e-01
-8.59762013e-01 4.51717734e-01 7.41994560e-01 -3.51327956e-01
-2.40799084e-01 -2.23734096e-01 1.07477911e-01 2.00195491e-01
6.71927691e-01 -2.03313641e-02 -4.43129778e-01 2.76125997e-01
-3.25466365e-01 4.17728484e-01 4.10633057e-01 9.99325812e-01
5.81231296e-01 1.79060385e-01 -1.04814701e-01 7.24639595e-01
6.22163355e-01 1.05098331e+00 -3.21075656e-02 -8.43187630e-01
6.46102369e-01 6.34118617e-01 4.29138511e-01 -9.24336791e-01
-5.59142292e-01 -5.26350319e-01 -7.76178420e-01 5.98905981e-01
2.44599521e-01 2.26863306e-02 -1.38314283e+00 1.42964172e+00
6.49073780e-01 2.62969643e-01 7.90655166e-02 1.55621397e+00
1.24031401e+00 7.91209280e-01 -6.98095113e-02 -3.97498190e-01
1.22496939e+00 -7.49600172e-01 -4.94273871e-01 -6.62972510e-01
2.41357312e-02 -7.52583265e-01 1.02418852e+00 3.69028687e-01
-8.55623126e-01 -6.98472977e-01 -1.18583584e+00 -6.44725859e-02
-3.21598023e-01 -1.37114644e-01 5.12692809e-01 7.75592744e-01
-8.76415253e-01 -2.18533173e-01 -4.42267418e-01 -2.85954297e-01
4.38508630e-01 -4.42582108e-02 -7.74365664e-02 -7.84684777e-01
-1.14372694e+00 1.01397002e+00 3.77348334e-01 4.89096284e-01
-1.30872500e+00 -7.31421709e-01 -8.43879700e-01 -2.99055815e-01
4.20107722e-01 -8.72935951e-01 9.50073242e-01 -2.91458994e-01
-1.12926900e+00 9.20170784e-01 -9.76985469e-02 -3.63040447e-01
5.69793463e-01 -6.27809465e-01 -4.71891463e-01 9.18962210e-02
1.51966251e-02 6.48690820e-01 1.13010025e+00 -1.46923459e+00
-9.76890743e-01 -6.57726943e-01 4.62886959e-01 6.98584259e-01
-1.80605724e-01 -7.26718130e-03 -6.42134249e-01 -2.57926881e-01
2.34908789e-01 -6.04927599e-01 -2.59288907e-01 4.25697893e-01
-1.69355705e-01 -1.39657125e-01 1.07941794e+00 -2.00735867e-01
5.98688304e-01 -2.03396058e+00 -1.91914827e-01 -4.98064548e-01
4.67291892e-01 3.84454668e-01 -9.75853056e-02 -2.25358918e-01
5.56007981e-01 -3.71875048e-01 -3.39129180e-01 -4.48132128e-01
-2.38894671e-01 4.06283438e-01 -2.90801734e-01 6.89930797e-01
3.59007120e-01 3.95672679e-01 -9.72831070e-01 -3.88901770e-01
6.14449680e-01 5.64431429e-01 -3.59358460e-01 4.94944274e-01
-2.05284700e-01 3.40805560e-01 -5.37368059e-01 1.25178075e+00
1.40752137e+00 3.02404106e-01 -6.22342408e-01 -4.58302870e-02
-3.08097929e-01 -3.03202588e-02 -1.16167545e+00 1.28292620e+00
-4.47557956e-01 4.90660340e-01 2.97467500e-01 -4.67786849e-01
1.15032077e+00 -8.94727111e-02 -7.51852542e-02 -9.45606768e-01
2.17084885e-01 2.10094243e-01 -2.52068937e-01 -8.13871443e-01
4.43954259e-01 -7.89204538e-02 1.51657328e-01 -6.20371737e-02
-2.93230265e-01 -4.54297245e-01 3.15563940e-02 1.42476112e-01
9.92342353e-01 -9.73380506e-02 7.59852165e-03 1.45285621e-01
3.98375183e-01 -2.46878907e-01 8.67069721e-01 1.05515683e+00
-9.23532426e-01 7.49702930e-01 3.55575792e-03 -4.31796521e-01
-7.19321847e-01 -1.44048810e+00 -3.31034303e-01 8.93714070e-01
8.77809227e-01 2.29419619e-01 -2.94236660e-01 -5.22996902e-01
1.30732492e-01 4.27106708e-01 -5.01859367e-01 -2.57586539e-01
-4.16216105e-01 -1.19594002e+00 3.21081817e-01 4.09981728e-01
1.19134009e+00 -9.11159992e-01 -8.45384598e-01 -3.74537408e-02
-1.16866417e-01 -1.55874634e+00 -2.45789438e-01 6.70765787e-02
-8.12534750e-01 -1.10879266e+00 -5.09455562e-01 -6.36669040e-01
4.65296984e-01 1.32631934e+00 1.08455849e+00 -7.22246245e-02
-6.18789196e-01 1.79042250e-01 -4.64765072e-01 -8.56852829e-01
-7.16641098e-02 -3.65568459e-01 3.06529224e-01 -7.10816309e-02
5.42128801e-01 -4.24148619e-01 -1.02167869e+00 3.22032154e-01
-8.36446524e-01 -1.36695758e-01 6.14374042e-01 6.73978567e-01
3.42129052e-01 3.35413665e-02 4.41309512e-01 -3.75640541e-01
1.87400207e-01 -5.56326389e-01 -9.04506624e-01 -2.29766220e-01
-5.15153348e-01 -6.97657943e-01 2.84071714e-01 -2.50710756e-01
-1.33686912e+00 1.37800783e-01 4.03695107e-02 -5.82585394e-01
-4.71320480e-01 -1.20180540e-01 -4.19522077e-01 -3.16578150e-01
7.95262098e-01 3.10871571e-01 -2.31621385e-01 -4.19584930e-01
2.09669679e-01 8.38607073e-01 6.41043127e-01 -1.03250332e-01
1.20405579e+00 7.37141550e-01 -8.23991522e-02 -9.33243513e-01
-1.29897225e+00 -6.98152483e-01 -2.96118259e-01 -5.74015498e-01
8.53269935e-01 -1.57607818e+00 -5.79414845e-01 9.91162956e-01
-1.07348573e+00 -6.72432035e-02 2.67519355e-01 5.12765110e-01
-6.14820085e-02 3.85920107e-01 -5.72560906e-01 -1.10636795e+00
-4.21951443e-01 -9.39523101e-01 1.11866784e+00 7.15797961e-01
6.02343321e-01 -4.43611830e-01 -3.15724134e-01 6.52306557e-01
4.06352937e-01 2.21806780e-01 5.18797755e-01 3.10046554e-01
-9.92089570e-01 -1.82315335e-01 -7.48981178e-01 5.44126928e-01
2.67895132e-01 -1.96706519e-01 -1.38755560e+00 -3.96672636e-01
2.38170102e-01 -3.23105156e-01 1.22400606e+00 4.38447893e-01
8.90517056e-01 2.47534811e-01 -1.24379270e-01 9.01457310e-01
1.43977606e+00 2.85701632e-01 4.61878002e-01 4.25925881e-01
9.07400370e-01 4.18314755e-01 9.25586939e-01 5.39392710e-01
4.14132804e-01 3.43465149e-01 1.22105765e+00 -2.26398647e-01
-2.89576679e-01 1.17255166e-01 3.69514644e-01 1.88524112e-01
5.90231940e-02 -4.06880915e-01 -6.24462128e-01 6.22954786e-01
-1.52784562e+00 -6.58268571e-01 -1.53629079e-01 1.90066493e+00
5.22535264e-01 5.33503175e-01 -1.59467012e-01 -7.72019476e-02
4.84924972e-01 4.19597566e-01 -1.06177950e+00 6.47141039e-02
-5.51357150e-01 -4.03870016e-01 6.11448944e-01 3.43030989e-01
-1.30709529e+00 8.35832655e-01 5.94926357e+00 2.81242937e-01
-1.06164122e+00 -7.57647082e-02 3.42405289e-01 -4.30893689e-01
3.39953378e-02 -4.31873649e-01 -1.07428575e+00 4.84890401e-01
2.21156001e-01 2.47144878e-01 1.47108331e-01 9.19731438e-01
5.99224985e-01 -3.58858705e-01 -5.99625885e-01 1.08798814e+00
2.70152926e-01 -8.11078072e-01 -3.11398692e-02 -1.79746866e-01
7.39437342e-01 4.48537320e-01 2.15500221e-01 2.27935091e-01
1.17416829e-01 -7.50173151e-01 6.39967442e-01 1.95984274e-01
6.14500105e-01 -5.29064476e-01 7.65798748e-01 5.04049063e-01
-1.03298879e+00 -2.73883134e-01 -6.84606671e-01 -4.54684854e-01
1.01846486e-01 9.85801339e-01 -7.90535510e-01 2.86565274e-01
1.30484712e+00 8.64208221e-01 -5.32545686e-01 1.37221396e+00
-3.40696126e-01 3.61829340e-01 -4.35080051e-01 3.14073116e-02
3.08265835e-01 1.59993954e-02 1.01479936e+00 1.11438632e+00
8.83207396e-02 3.63659143e-01 2.37458214e-01 6.86496556e-01
1.15263658e-02 -5.18750489e-01 -7.78242111e-01 7.26082206e-01
5.22709787e-01 1.15157592e+00 -4.02241886e-01 -1.71285346e-01
-5.01665175e-01 8.73467624e-01 3.71433087e-02 5.04490674e-01
-8.70613217e-01 -2.63095587e-01 1.08913600e+00 -7.64292926e-02
3.47315341e-01 -3.83035332e-01 -2.57820725e-01 -1.17860186e+00
3.72666419e-01 -5.41172802e-01 1.64285854e-01 -9.43089485e-01
-1.37160814e+00 4.34199363e-01 -7.45260790e-02 -1.27822816e+00
3.44096810e-01 -8.80649447e-01 -6.23589516e-01 8.88347208e-01
-2.10808682e+00 -8.60455096e-01 -9.80991423e-01 6.12240851e-01
8.38403404e-01 7.87316412e-02 2.48582870e-01 5.92610598e-01
-7.75851786e-01 3.74955028e-01 -2.35400703e-02 -3.49804819e-01
8.63850296e-01 -1.18517482e+00 4.28446174e-01 1.21054018e+00
-3.31226319e-01 7.34530911e-02 8.27828586e-01 -4.38960791e-01
-1.58188045e+00 -1.43241096e+00 4.42566872e-01 -4.52512532e-01
1.92068547e-01 -4.81050313e-01 -9.62997794e-01 1.77967519e-01
-3.58410478e-01 4.43310738e-01 -9.98131633e-02 -2.80458778e-01
-5.13349295e-01 -5.00129402e-01 -1.27392197e+00 4.84807163e-01
1.32002461e+00 -2.64310956e-01 -5.36244750e-01 3.83421570e-01
1.08659101e+00 -7.74040818e-01 -2.98889637e-01 8.40778708e-01
3.83946687e-01 -1.24783206e+00 1.12576497e+00 -6.06658198e-02
1.62088484e-01 -8.27373683e-01 -4.37841952e-01 -1.09297466e+00
-2.42994979e-01 -1.53183997e-01 -5.21149099e-01 9.32699144e-01
-9.12052300e-03 -6.54461861e-01 8.10742319e-01 9.97441411e-02
-5.29120564e-01 -6.32604122e-01 -9.50926900e-01 -6.15081072e-01
-2.80479163e-01 -4.43566024e-01 3.17754686e-01 6.29447401e-01
-9.10350561e-01 1.89701021e-01 -4.74804878e-01 8.31166387e-01
1.02078414e+00 2.84624100e-01 8.22870791e-01 -1.08559024e+00
-4.65956405e-02 -1.54468045e-01 -5.35842419e-01 -1.42706883e+00
-3.68199259e-01 -2.50383586e-01 5.62602878e-01 -1.63705230e+00
2.01109573e-01 -4.71209645e-01 -4.07581747e-01 1.18539467e-01
-5.87518692e-01 5.01343429e-01 1.99283600e-01 2.49024466e-01
-7.27305889e-01 6.50395989e-01 1.52373195e+00 -2.77165294e-01
-1.38218239e-01 1.29346162e-01 -5.83981335e-01 8.61944616e-01
7.28741527e-01 -1.80485368e-01 -5.12916207e-01 -9.20478225e-01
-8.36507455e-02 -2.79268891e-01 6.64182007e-01 -1.19929302e+00
9.77682546e-02 -2.45401338e-01 7.97234535e-01 -9.50318217e-01
4.64725137e-01 -6.42139018e-01 -5.37332654e-01 4.75593805e-01
3.83163601e-01 -3.55040222e-01 4.25339758e-01 9.57932115e-01
-1.51671141e-01 1.26230597e-01 1.02441299e+00 -2.44146883e-01
-1.27846932e+00 4.91081953e-01 -2.45626315e-01 -1.11306019e-01
1.03260469e+00 -5.44725716e-01 -7.03543544e-01 -1.38294986e-02
-2.26523116e-01 6.29269421e-01 2.58681178e-01 7.57517338e-01
9.71510410e-01 -8.46819520e-01 -8.96885395e-01 3.59344423e-01
4.11994219e-01 5.27449012e-01 6.17215216e-01 5.25303543e-01
-5.50426960e-01 1.66626215e-01 -1.38700202e-01 -8.96129191e-01
-1.13394070e+00 3.41922879e-01 6.38214231e-01 3.15117806e-01
-7.33936727e-01 1.13214064e+00 6.99107349e-01 -1.13397971e-01
5.14029622e-01 -3.95088911e-01 -5.64943776e-02 -1.70795992e-01
7.71245122e-01 2.78155148e-01 9.98130813e-02 -2.64986604e-01
-4.08245981e-01 6.68525159e-01 -3.55620950e-01 6.65471673e-01
1.09358549e+00 -5.82543790e-01 2.46735960e-01 3.22145969e-01
7.30931997e-01 -4.43125486e-01 -1.93614769e+00 -3.64387900e-01
-7.83660889e-01 -8.45998824e-01 4.16147441e-01 -8.42440248e-01
-1.03113496e+00 8.58564615e-01 1.18382406e+00 2.54796389e-02
1.51107407e+00 -8.29271227e-02 8.22420478e-01 4.04215068e-01
4.33245808e-01 -6.46914303e-01 1.83736652e-01 8.02373886e-01
6.18192613e-01 -1.91112459e+00 -2.94218626e-04 -5.52995145e-01
-2.19722867e-01 7.64020741e-01 1.25188994e+00 -1.12271093e-01
4.44307625e-01 1.73514903e-01 3.52936774e-01 -1.90658495e-01
-6.21484756e-01 -4.30886030e-01 -8.05835426e-03 8.84812832e-01
-1.85712263e-01 -5.83291054e-02 2.65882939e-01 1.52513802e-01
1.71158433e-01 -3.14881086e-01 5.16761243e-01 8.71076405e-01
-1.04234886e+00 -1.96809158e-01 -6.67316079e-01 3.98409337e-01
-1.44991934e-01 -1.12617843e-01 -2.34033212e-01 7.14733779e-01
6.52547121e-01 1.17515147e+00 1.47909999e-01 -3.23515445e-01
5.94552875e-01 -6.55463755e-01 5.04820049e-01 -5.30309558e-01
-1.43991793e-02 -1.27289891e-01 -1.45719394e-01 -4.22947556e-01
-5.02093315e-01 -4.23800588e-01 -8.44078243e-01 -1.93203047e-01
-5.22572339e-01 -5.63510716e-01 4.87580001e-01 8.48720491e-01
1.99217260e-01 4.89882678e-01 7.56406844e-01 -1.16899407e+00
-3.27485263e-01 -1.03612399e+00 -7.04550266e-01 1.48432059e-02
8.83297622e-01 -1.01546967e+00 -7.64817357e-01 -1.96960881e-01]
|
[8.183815002441406, -2.369802713394165]
|
c2519f61-f9aa-48dc-91f7-a823185e18da
|
recursive-metropolis-hastings-naming-game
|
2305.19761
| null |
https://arxiv.org/abs/2305.19761v1
|
https://arxiv.org/pdf/2305.19761v1.pdf
|
Recursive Metropolis-Hastings Naming Game: Symbol Emergence in a Multi-agent System based on Probabilistic Generative Models
|
In the studies on symbol emergence and emergent communication in a population of agents, a computational model was employed in which agents participate in various language games. Among these, the Metropolis-Hastings naming game (MHNG) possesses a notable mathematical property: symbol emergence through MHNG is proven to be a decentralized Bayesian inference of representations shared by the agents. However, the previously proposed MHNG is limited to a two-agent scenario. This paper extends MHNG to an N-agent scenario. The main contributions of this paper are twofold: (1) we propose the recursive Metropolis-Hastings naming game (RMHNG) as an N-agent version of MHNG and demonstrate that RMHNG is an approximate Bayesian inference method for the posterior distribution over a latent variable shared by agents, similar to MHNG; and (2) we empirically evaluate the performance of RMHNG on synthetic and real image data, enabling multiple agents to develop and share a symbol system. Furthermore, we introduce two types of approximations -- one-sample and limited-length -- to reduce computational complexity while maintaining the ability to explain communication in a population of agents. The experimental findings showcased the efficacy of RMHNG as a decentralized Bayesian inference for approximating the posterior distribution concerning latent variables, which are jointly shared among agents, akin to MHNG. Moreover, the utilization of RMHNG elucidated the agents' capacity to exchange symbols. Furthermore, the study discovered that even the computationally simplified version of RMHNG could enable symbols to emerge among the agents.
|
['Yoshinobu Hagiwara', 'Akira Taniguchi', 'Tadahiro Taniguchi', 'Jun Inukai']
|
2023-05-31
| null | null | null | null |
['bayesian-inference']
|
['methodology']
|
[-2.03601018e-01 4.67694908e-01 3.23866338e-01 3.43759775e-01
6.21505417e-02 -2.12704748e-01 1.09653366e+00 -2.71005780e-01
-2.51992106e-01 9.11753774e-01 -1.12679370e-01 -3.13959390e-01
-4.11503702e-01 -1.00939381e+00 -5.99729538e-01 -9.67299163e-01
-6.96169436e-01 7.63767183e-01 5.52591991e-05 -1.20763376e-01
4.15641218e-02 4.43764359e-01 -1.57210803e+00 -5.61047137e-01
7.60769069e-01 4.05881315e-01 5.23411274e-01 9.02933121e-01
2.99581200e-01 1.08933675e+00 -9.28999782e-01 -7.48071969e-02
4.35723752e-01 -7.58643866e-01 -4.74843651e-01 2.51101851e-01
-4.15393114e-01 -7.05849349e-01 -4.56489801e-01 1.02915597e+00
2.46933058e-01 -1.38460537e-02 9.34643030e-01 -1.70258617e+00
-8.05546880e-01 1.07680929e+00 -3.82831067e-01 -2.44825587e-01
2.15455264e-01 2.22543895e-01 6.68431640e-01 -5.16293645e-01
5.95113158e-01 1.26880622e+00 3.95593882e-01 3.49116176e-01
-1.08542955e+00 -8.66406262e-01 1.13939475e-02 -3.59513611e-02
-2.00926304e+00 -1.95809722e-01 3.75717878e-01 -4.50424433e-01
9.85861599e-01 7.64795318e-02 1.17791533e+00 8.83597910e-01
4.72597361e-01 7.66732574e-01 1.34104908e+00 -5.98075747e-01
7.15938568e-01 -1.16491988e-01 -2.90176213e-01 8.08609426e-01
6.04328811e-01 2.93660462e-01 -6.42282367e-01 -6.91830754e-01
1.44619298e+00 -2.13461787e-01 -2.18425300e-02 -1.60622716e-01
-1.27694094e+00 9.61605370e-01 -1.80106014e-01 8.09676796e-02
-7.50718713e-01 5.28467774e-01 -1.60937116e-01 4.64226931e-01
3.72167200e-01 2.37696454e-01 8.85116830e-02 -2.76239872e-01
-7.24821448e-01 4.78891760e-01 1.27328551e+00 1.14586413e+00
7.07733929e-01 9.54061225e-02 1.08185224e-01 2.75271714e-01
7.56308615e-01 1.12586546e+00 2.05514073e-01 -1.23418379e+00
-1.61863700e-01 9.48543176e-02 2.59070486e-01 -8.77847254e-01
-2.79057264e-01 -4.55685377e-01 -1.13887501e+00 2.38177255e-01
3.79716039e-01 -2.32348964e-01 -4.69595641e-01 2.15252852e+00
1.49676666e-01 4.97243553e-01 4.88641053e-01 4.18139368e-01
3.11745644e-01 8.22644353e-01 -2.99901962e-01 -6.58368111e-01
1.29032660e+00 -7.09957719e-01 -7.92267501e-01 4.04190958e-01
4.27117229e-01 -4.07252908e-01 4.63406563e-01 4.98618245e-01
-1.24956858e+00 -2.08688602e-01 -7.61340201e-01 8.00591230e-01
-1.24632917e-01 -3.98566067e-01 8.46243083e-01 6.29154444e-01
-1.52968931e+00 1.12533621e-01 -9.66411591e-01 -6.65787578e-01
1.03195898e-01 5.19089341e-01 4.14992087e-02 3.87360752e-01
-1.13462579e+00 6.48032784e-01 3.16137880e-01 -1.14230528e-01
-1.34765923e+00 -4.45790105e-02 -4.72051233e-01 8.25202614e-02
3.70241195e-01 -1.22892106e+00 1.28623569e+00 -8.62235904e-01
-2.12557387e+00 3.05385649e-01 -1.30055040e-01 -7.19844818e-01
5.20074666e-01 2.54906356e-01 -1.38354003e-01 1.10100426e-01
1.77277833e-01 6.37090206e-01 8.08855712e-01 -1.42386055e+00
-4.05426085e-01 1.24719217e-01 1.09942324e-01 2.22208932e-01
-7.84349963e-02 -2.23870873e-01 -1.86518475e-01 -6.35803044e-01
-8.89766440e-02 -1.08746409e+00 -1.54461339e-01 -3.64754558e-01
-4.29746509e-01 -5.15289187e-01 5.83882511e-01 -2.03711465e-01
1.01119196e+00 -2.07440352e+00 3.38795155e-01 5.05259812e-01
5.53584218e-01 -1.72572732e-01 -1.91638365e-01 1.12484288e+00
6.18370056e-01 -4.23391685e-02 -1.71705186e-01 -5.60413897e-01
2.93507993e-01 5.21636963e-01 -3.92663479e-01 5.58565199e-01
-4.08303469e-01 9.34100568e-01 -9.57018316e-01 -4.17163432e-01
1.08509660e-01 3.01024050e-01 -4.74681199e-01 1.37537301e-01
-1.46607503e-01 3.81086707e-01 -3.84408325e-01 2.53897667e-01
5.65691054e-01 -6.58912539e-01 4.28767264e-01 6.12115860e-01
-2.35414878e-01 -3.25837940e-01 -1.22358108e+00 1.27278483e+00
-3.73909138e-02 5.81550539e-01 9.58349630e-02 -5.29020488e-01
8.35277319e-01 4.54361647e-01 3.96449149e-01 -2.23503381e-01
3.47381443e-01 2.48027340e-01 1.31794721e-01 -2.74033964e-01
4.53760713e-01 2.61380542e-02 -2.06441045e-01 1.08354664e+00
-3.13397437e-01 -3.24261457e-01 3.40086311e-01 7.86281109e-01
1.12044024e+00 -2.52528012e-01 7.17237234e-01 -3.91071111e-01
1.24791116e-01 -4.46527660e-01 2.36030146e-01 1.71888840e+00
-3.81580810e-03 -1.83288772e-02 3.40393811e-01 -1.72985420e-01
-1.14640117e+00 -1.43759203e+00 1.86097801e-01 6.63893878e-01
5.09450853e-01 -5.84909558e-01 -9.41029727e-01 7.60683045e-02
-6.22981880e-03 7.80356467e-01 -5.21092594e-01 9.42413211e-02
-4.69394550e-02 -8.50647569e-01 8.30269992e-01 -7.51695335e-02
8.21595192e-01 -1.10587597e+00 -1.11183214e+00 2.70188928e-01
-6.15021177e-02 -8.87222946e-01 -8.18284303e-02 -1.82223842e-01
-4.16164130e-01 -8.61855567e-01 -6.74005449e-01 -4.87834483e-01
8.53598833e-01 1.95443228e-01 7.18401015e-01 1.95942953e-01
3.22895870e-02 8.60361397e-01 -4.82707560e-01 -4.94707823e-01
-8.46799612e-01 3.09862141e-02 5.86699009e-01 2.04969849e-02
3.50668803e-02 -8.30937684e-01 -5.04637420e-01 2.46827856e-01
-1.00605524e+00 5.55504918e-01 7.69989133e-01 6.29870355e-01
-1.12816364e-01 2.86560208e-01 5.97226501e-01 -4.49572802e-01
1.02323902e+00 -5.02565324e-01 -8.36062312e-01 1.85635477e-01
-5.16280711e-01 -1.75095573e-02 5.54767966e-01 -4.86575425e-01
-7.14392483e-01 -3.43761772e-01 4.97989684e-01 4.94220294e-03
-8.69570524e-02 5.67588568e-01 4.39912140e-01 1.29215091e-01
7.67069533e-02 7.76262820e-01 5.63920081e-01 -4.61971983e-02
3.68985325e-01 7.74459362e-01 1.85837880e-01 -6.98728204e-01
9.06348944e-01 6.12586796e-01 2.32443944e-01 -1.28974605e+00
2.28548888e-02 2.21602663e-01 -1.43586561e-01 -2.67663985e-01
5.83109558e-01 -9.17131782e-01 -1.31297374e+00 1.11826456e+00
-1.37059236e+00 -7.27658689e-01 -6.04341984e-01 7.31912196e-01
-8.52339506e-01 5.79862773e-01 -8.31467450e-01 -1.39392126e+00
7.00400397e-02 -1.08441043e+00 7.03341246e-01 2.99051285e-01
-1.70323312e-01 -9.43158507e-01 3.92998874e-01 -9.84976366e-02
6.82152212e-01 -1.25610873e-01 8.64202976e-01 -5.26262879e-01
-1.07853401e+00 2.08224982e-01 4.71879616e-02 -3.02770227e-01
1.02583371e-01 1.62782177e-01 -4.47830677e-01 -7.41932213e-01
-7.15798810e-02 -1.84562862e-01 2.35496715e-01 7.51161218e-01
3.08594882e-01 -5.17717004e-01 -2.90038943e-01 2.81931430e-01
1.02517295e+00 3.07361394e-01 4.48389053e-01 3.32761228e-01
7.59149045e-02 4.02215481e-01 -3.52126211e-02 1.11040330e+00
7.35185266e-01 7.04780459e-01 4.35353458e-01 3.28742228e-02
2.08218709e-01 -1.92342222e-01 5.91387749e-01 1.47743738e+00
-3.44458371e-01 -6.06705785e-01 -6.73694611e-01 2.53018022e-01
-2.22606039e+00 -1.00484896e+00 8.51326361e-02 1.96964741e+00
4.90740836e-01 -3.09232324e-01 2.12949961e-01 -4.54201281e-01
7.88889110e-01 -9.35918987e-02 -5.33757091e-01 9.92564391e-03
-4.57318157e-01 -2.01713338e-01 4.03341830e-01 6.82989895e-01
-4.94185150e-01 9.66559231e-01 6.80902290e+00 8.25589120e-01
-5.05128384e-01 1.81605101e-01 1.92423910e-01 1.49057940e-01
-9.34772491e-02 2.73814112e-01 -6.93897724e-01 5.39863229e-01
9.45027351e-01 -4.39012855e-01 1.05446088e+00 1.86143667e-01
5.11676550e-01 -3.52434874e-01 -7.06971943e-01 9.79210734e-01
-1.16622839e-02 -1.19035709e+00 3.05627823e-01 7.73408353e-01
9.89966154e-01 8.32342431e-02 8.91463012e-02 3.36451791e-02
1.10433805e+00 -9.52676773e-01 1.01666284e+00 7.85497069e-01
2.82475948e-01 -6.31235480e-01 7.35511780e-01 7.88909912e-01
-1.23946655e+00 -9.63405818e-02 -3.16044211e-01 -6.16312623e-01
5.14897346e-01 3.00379455e-01 -9.05302227e-01 5.67756534e-01
3.23958009e-01 2.84635663e-01 -2.49464199e-01 8.08688104e-01
-2.63340265e-01 5.04468024e-01 -7.17616618e-01 -4.14947987e-01
3.15150201e-01 -4.83123690e-01 8.20775032e-01 7.38695383e-01
6.26848876e-01 1.79136992e-01 3.69215280e-01 1.06374991e+00
3.22173983e-01 -1.37608007e-01 -7.33875513e-01 -6.38405606e-02
1.02346361e+00 7.54522145e-01 -9.97531593e-01 -4.92407382e-01
-2.08382532e-01 8.61683846e-01 1.29164815e-01 6.28816366e-01
-9.04806256e-01 -1.73426807e-01 3.16931695e-01 -8.60114023e-02
4.05320287e-01 -5.37617862e-01 3.64621192e-01 -1.08809543e+00
-5.75574875e-01 -8.63919616e-01 -1.04952827e-01 -8.15320611e-01
-1.27736354e+00 5.35931826e-01 6.16485596e-01 -8.91238868e-01
-6.66753054e-01 -1.50921404e-01 -2.31624290e-01 5.49638033e-01
-9.92430449e-01 -1.12401259e+00 -3.11165303e-02 7.23345637e-01
3.83505940e-01 -6.77647173e-01 8.68179023e-01 -3.62852752e-01
-4.35536385e-01 2.43385613e-01 5.51821172e-01 -8.29245001e-02
-8.71374011e-02 -1.00281632e+00 2.77277410e-01 7.40372896e-01
2.70966291e-01 9.78548467e-01 8.69784415e-01 -8.45890522e-01
-1.32618797e+00 -6.35755181e-01 4.77436185e-01 9.68082994e-02
6.66695476e-01 -4.85234439e-01 -3.91758233e-01 8.73588681e-01
6.88140213e-01 -5.40920615e-01 5.99356949e-01 -2.83437371e-01
1.08338982e-01 3.74724865e-01 -8.26795459e-01 1.15902150e+00
9.31648016e-01 -5.50296605e-01 -3.45069587e-01 1.90551072e-01
4.12418336e-01 1.04091033e-01 -6.64651692e-01 -1.53331801e-01
6.20234489e-01 -9.77173328e-01 5.33761978e-01 1.21261999e-02
-6.09244704e-02 -6.78794622e-01 -2.03494325e-01 -1.18039656e+00
-3.58680129e-01 -1.25319421e+00 -2.87416726e-01 9.99099135e-01
7.53909945e-02 -1.41491008e+00 5.53030849e-01 4.20376807e-01
3.08443606e-01 -4.13428485e-01 -1.14165735e+00 -1.07503092e+00
-1.88688468e-02 -1.39196336e-01 8.53608549e-01 6.91972196e-01
9.26765203e-02 -1.89790428e-02 -6.88566446e-01 3.91003251e-01
1.06265759e+00 -1.72632426e-01 1.34532499e+00 -1.13694334e+00
-8.23548317e-01 -3.12080383e-01 -3.37177485e-01 -1.54862690e+00
2.45226726e-01 -7.60527551e-01 8.13950747e-02 -1.27863359e+00
6.91916525e-01 -2.57014185e-01 1.22259534e-03 1.00161426e-01
3.99450153e-01 -4.35632914e-02 3.93469125e-01 7.91938663e-01
-7.89271235e-01 8.76107872e-01 1.08320999e+00 4.25667912e-02
-2.15381622e-01 -1.66367650e-01 -2.95476645e-01 6.68112278e-01
5.99108338e-01 -5.88493049e-01 -6.80500507e-01 -4.98394743e-02
3.83081049e-01 2.37242922e-01 6.08578682e-01 -7.50450253e-01
7.68960834e-01 -1.03226215e-01 -3.10508519e-01 -5.90730369e-01
4.04085726e-01 -6.33805394e-01 9.90029216e-01 9.31744814e-01
-1.01976164e-01 1.63126066e-01 -2.98331052e-01 7.28363454e-01
2.18579143e-01 -2.78234005e-01 4.14152741e-01 -2.39892736e-01
-2.87951678e-01 -1.46942809e-01 -1.43824637e+00 -2.17502132e-01
1.28789544e+00 -3.57794076e-01 -4.35346425e-01 -1.00061953e+00
-5.03621280e-01 2.63995752e-02 7.09501326e-01 -2.76945651e-01
6.89311326e-01 -1.04325390e+00 -7.85169005e-01 3.61178637e-01
-3.13981324e-01 -9.90290418e-02 2.85645877e-03 8.92104626e-01
-7.53496885e-01 3.58758539e-01 -9.66008566e-03 -7.30635345e-01
-8.86952221e-01 3.03325325e-01 1.77837595e-01 -4.74982589e-01
-6.37913764e-01 5.86739242e-01 4.27810609e-01 -3.56096655e-01
-4.59674299e-02 -1.19414337e-01 2.82292455e-01 -4.33696181e-01
2.40847781e-01 4.24306691e-01 -1.03865385e+00 -6.39566481e-01
-9.65471268e-02 4.66231138e-01 -1.09741196e-01 -6.48884594e-01
1.23124123e+00 -6.05398893e-01 -4.82790470e-01 5.83338618e-01
3.72184455e-01 -9.06077102e-02 -1.30758393e+00 -2.93806911e-01
-4.47318405e-01 -2.07807332e-01 -2.33887613e-01 -3.11012655e-01
-6.53876901e-01 8.14256221e-02 2.45044440e-01 9.03076768e-01
7.81002283e-01 2.36352369e-01 3.23085576e-01 5.14327884e-01
1.06412005e+00 -8.61616015e-01 3.86679396e-02 5.57419837e-01
4.88386422e-01 -6.63827419e-01 1.07878074e-01 -2.18395531e-01
-3.37416202e-01 9.99529421e-01 3.24512944e-02 -5.20231538e-02
7.31961370e-01 4.16253924e-01 -1.97439417e-01 -5.15335500e-01
-8.81037056e-01 -9.41603817e-03 -6.75454199e-01 5.66311181e-01
-1.72001511e-01 3.04657310e-01 -2.11432606e-01 2.99985230e-01
-3.92877191e-01 1.70025319e-01 1.00745237e+00 1.00487316e+00
-3.11671346e-01 -9.27571952e-01 -3.58380228e-01 1.66766286e-01
9.71164480e-02 -1.84038863e-01 -2.68376142e-01 9.73684430e-01
-3.30845788e-02 1.17685616e+00 1.46486357e-01 -1.17060773e-01
-2.38111749e-01 -3.58629316e-01 4.87693936e-01 -3.03407758e-01
-3.84858608e-01 9.65237692e-02 -3.16696256e-01 -2.87216268e-02
-7.27341890e-01 -7.99128234e-01 -9.91418362e-01 -7.43454933e-01
-6.31962478e-01 6.52373850e-01 2.26669475e-01 1.07523263e+00
4.08476532e-01 2.33182475e-01 5.21464705e-01 -8.88762236e-01
-6.98982656e-01 -9.33896661e-01 -1.27027440e+00 7.61510432e-02
5.87341413e-02 -6.87810421e-01 -6.64954960e-01 2.53687724e-02]
|
[4.375793933868408, 2.77817964553833]
|
027576fe-619c-4344-85f0-f8981db1a1a6
|
adaptive-energy-management-for-real-driving
|
2007.12560
| null |
https://arxiv.org/abs/2007.12560v1
|
https://arxiv.org/pdf/2007.12560v1.pdf
|
Adaptive Energy Management for Real Driving Conditions via Transfer Reinforcement Learning
|
This article proposes a transfer reinforcement learning (RL) based adaptive energy managing approach for a hybrid electric vehicle (HEV) with parallel topology. This approach is bi-level. The up-level characterizes how to transform the Q-value tables in the RL framework via driving cycle transformation (DCT). Especially, transition probability matrices (TPMs) of power request are computed for different cycles, and induced matrix norm (IMN) is employed as a critical criterion to identify the transformation differences and to determine the alteration of the control strategy. The lower-level determines how to set the corresponding control strategies with the transformed Q-value tables and TPMs by using model-free reinforcement learning (RL) algorithm. Numerical tests illustrate that the transferred performance can be tuned by IMN value and the transfer RL controller could receive a higher fuel economy. The comparison demonstrates that the proposed strategy exceeds the conventional RL approach in both calculation speed and control performance.
|
['Wenhao Tan', 'Xiaolin Tang', 'Teng Liu', 'Jiaxin Chen', 'Dongpu Cao']
|
2020-07-24
| null | null | null | null |
['transfer-reinforcement-learning']
|
['methodology']
|
[ 9.12894960e-03 9.01076943e-02 -4.78532732e-01 1.95641086e-01
-4.02472645e-01 -4.92298603e-01 6.35717988e-01 1.53273651e-02
-2.74843305e-01 1.12910116e+00 -3.97580355e-01 -2.17043847e-01
-6.88484132e-01 -9.64165568e-01 -5.17679393e-01 -1.01712132e+00
-1.28952460e-02 2.49442443e-01 8.48394409e-02 -4.85218763e-01
4.85695213e-01 6.54820681e-01 -1.37755084e+00 -3.46643209e-01
8.73600602e-01 8.98760915e-01 2.13562369e-01 5.07270038e-01
1.20108224e-01 8.46209645e-01 -4.29659039e-01 2.80637950e-01
1.64900064e-01 -6.47657275e-01 -5.31339586e-01 -1.45381853e-01
-8.46940935e-01 -8.52869824e-02 -1.23561881e-01 1.07243371e+00
5.55014729e-01 4.32935148e-01 9.91991460e-01 -1.93316782e+00
3.14233959e-01 5.11253893e-01 -5.13161361e-01 3.13897014e-01
5.74414879e-02 3.64831030e-01 4.64045763e-01 -6.02347910e-01
2.26958543e-01 1.18198466e+00 4.14251566e-01 1.67931780e-01
-1.06996822e+00 -8.73909295e-01 -2.19675481e-01 8.45474482e-01
-1.44869006e+00 1.84173156e-02 1.01798403e+00 -3.07167590e-01
1.01304448e+00 2.58999407e-01 1.17953801e+00 1.71530992e-01
9.14898276e-01 3.52557063e-01 1.31077218e+00 -6.33031547e-01
5.42898238e-01 2.32055902e-01 -3.42299879e-01 4.56264913e-01
2.78782010e-01 3.11640263e-01 -2.29917206e-02 9.05660018e-02
4.73043293e-01 -5.47262251e-01 1.58145100e-01 -6.01936281e-01
-8.49756479e-01 8.96720171e-01 1.25791766e-02 3.73376757e-01
-4.28341627e-01 2.09495828e-01 6.14124060e-01 3.07342380e-01
-3.14246863e-01 4.77877676e-01 -3.96247029e-01 -8.92798081e-02
-4.58238840e-01 1.57056063e-01 6.37175322e-01 9.02595103e-01
9.48629975e-01 7.24117875e-01 -3.03587347e-01 4.62952316e-01
2.59335190e-01 6.56250000e-01 4.32657957e-01 -1.07147372e+00
2.47448727e-01 7.04426289e-01 3.13681036e-01 -7.71663845e-01
-6.00868523e-01 -1.16103612e-01 -7.43804932e-01 4.86720741e-01
1.03028543e-01 -6.74299061e-01 -2.32858226e-01 1.42413199e+00
4.94009256e-01 -2.26241097e-01 2.42421031e-01 4.12580937e-01
1.34604827e-01 1.20955336e+00 1.14396505e-01 -8.88596475e-01
1.21794772e+00 -6.11044645e-01 -1.23078859e+00 1.88166901e-01
5.45370460e-01 -4.41683412e-01 6.78940892e-01 2.76582897e-01
-1.00716329e+00 -5.19040048e-01 -1.54611099e+00 6.73065364e-01
-4.69391763e-01 2.65921324e-01 -5.55494502e-02 6.32558167e-01
-8.51450503e-01 7.22522378e-01 -6.02617443e-01 -2.65040040e-01
-2.68039525e-01 5.68814158e-01 3.16613227e-01 6.55720353e-01
-1.70137048e+00 1.33582437e+00 7.90046513e-01 1.88449651e-01
-6.67784333e-01 -8.16664457e-01 -6.20546758e-01 -2.07699295e-02
3.29382360e-01 -2.67998368e-01 9.93951857e-01 -8.77610862e-01
-2.26115322e+00 -4.93125580e-02 4.96975541e-01 -3.01940650e-01
4.83218640e-01 5.25643229e-01 -6.79577112e-01 2.76461720e-01
-2.41358399e-01 4.00190473e-01 1.03119886e+00 -1.14175224e+00
-8.84862661e-01 8.69892687e-02 -3.15323979e-01 5.96696615e-01
-2.03181639e-01 -3.54687512e-01 2.90348202e-01 -3.18390608e-01
-3.99704009e-01 -8.19601536e-01 -9.22478139e-02 -3.50546300e-01
-4.48418334e-02 -5.83929300e-01 9.84502316e-01 -4.74587977e-01
1.34089887e+00 -1.76255286e+00 2.33732551e-01 7.05648422e-01
-3.86697769e-01 9.86487791e-02 2.15534210e-01 7.02713728e-01
4.84460555e-02 -3.22768129e-02 1.25989942e-02 8.97163808e-01
1.37369588e-01 1.67709425e-01 1.14268228e-01 2.99991429e-01
3.22590619e-01 5.69206178e-01 -8.22096109e-01 -6.23985231e-01
5.40925980e-01 3.30369711e-01 -4.37402755e-01 2.27566838e-01
-2.54153349e-02 3.20946872e-01 -9.11247849e-01 -5.49423881e-03
5.06194055e-01 3.87385130e-01 6.32584929e-01 -7.21189618e-01
-4.69987631e-01 -4.42120254e-01 -1.52433789e+00 6.76611781e-01
-5.63358605e-01 4.52652782e-01 1.67999238e-01 -1.23015857e+00
1.32144296e+00 1.14297815e-01 9.19751585e-01 -9.95975316e-01
5.72260201e-01 4.90882769e-02 1.06388576e-01 -4.50108111e-01
3.43569726e-01 3.94323515e-03 -1.61124811e-01 2.10921720e-01
-1.17534921e-01 -6.20492816e-01 3.86301875e-01 -3.17154765e-01
6.80359364e-01 1.58304706e-01 6.17959321e-01 -8.68460357e-01
1.35288334e+00 1.74854174e-01 7.20365345e-01 1.01000085e-01
-2.19362572e-01 -8.28340411e-01 7.48747647e-01 -7.97945037e-02
-1.16648555e+00 -8.77514362e-01 -1.76938698e-01 5.84892154e-01
4.87760514e-01 2.52947390e-01 -7.64667630e-01 -1.59486815e-01
-1.10994793e-01 1.23934531e+00 -1.84893131e-01 -6.64344966e-01
-5.88675201e-01 -6.66800380e-01 1.07760906e-01 3.48200709e-01
7.01979876e-01 -9.14180994e-01 -1.02692699e+00 3.43754590e-01
9.67011824e-02 -7.70910740e-01 -2.60853380e-01 4.41855460e-01
-8.56124461e-01 -1.01602161e+00 -1.59826234e-01 -7.41535306e-01
7.06623554e-01 -4.17748600e-01 6.17743731e-01 -9.51752439e-02
-1.62879184e-01 4.36415046e-01 -2.10122362e-01 -3.21960300e-01
-8.20799947e-01 -2.14629099e-02 2.24464029e-01 -7.64478445e-02
4.75139730e-02 -2.51250505e-01 -5.00744104e-01 5.30171692e-01
-6.25665069e-01 -7.37379789e-02 5.74192762e-01 8.24141264e-01
8.49920511e-01 7.66096473e-01 1.16273880e+00 -1.59463301e-01
8.53607416e-01 -2.88364887e-01 -1.28040266e+00 3.24453741e-01
-1.26486444e+00 4.77092773e-01 9.82559621e-01 -2.87484795e-01
-1.06111503e+00 4.01337370e-02 1.67267337e-01 -2.30679587e-01
4.65769529e-01 -2.05422699e-01 -4.88660812e-01 -1.90413296e-01
-3.18831056e-02 4.18979347e-01 3.55896026e-01 2.11997971e-01
1.87171012e-01 4.30553406e-01 2.67923117e-01 -5.94802856e-01
1.08476257e+00 -3.22927326e-01 5.53478122e-01 -4.33855087e-01
3.07289481e-01 1.95998847e-01 -6.09398305e-01 -7.22040355e-01
8.44865263e-01 -5.80129445e-01 -1.29142320e+00 3.65714371e-01
-6.93493545e-01 -4.20333892e-01 -3.55877817e-01 5.13650715e-01
-1.07393408e+00 5.43377269e-03 -3.58761251e-01 -9.30086195e-01
-5.13591528e-01 -1.29689813e+00 2.31521383e-01 5.12835979e-01
7.10559115e-02 -9.94529605e-01 1.19633172e-02 -1.59018546e-01
3.43931735e-01 3.08567524e-01 1.56459653e+00 -2.06581071e-01
-3.55781019e-01 2.17458799e-01 2.94934988e-01 2.87419528e-01
3.33974920e-02 1.83410078e-01 -3.19033980e-01 -6.56112373e-01
-1.08729355e-01 1.02792226e-01 -2.09997654e-01 4.18384284e-01
1.01314723e+00 -4.79106814e-01 -4.22981918e-01 3.23429555e-01
1.97184229e+00 1.28687084e+00 7.92223334e-01 6.24206781e-01
2.90754139e-01 3.66005361e-01 9.20998931e-01 6.40785098e-01
2.01362431e-01 6.48899555e-01 4.02609080e-01 -1.33371039e-03
1.89026907e-01 -1.91100463e-01 4.93743002e-01 9.11392272e-01
3.74439992e-02 1.93168949e-02 -5.74000180e-01 1.81512579e-01
-1.54126394e+00 -8.94289196e-01 1.98531896e-01 2.11969781e+00
7.62603283e-01 2.60732383e-01 1.82571337e-01 6.51528180e-01
9.87680674e-01 -1.92100853e-01 -8.73971462e-01 -1.16339707e+00
3.40749174e-01 -4.41430658e-02 8.21928144e-01 4.50305104e-01
-4.66350496e-01 2.82352984e-01 6.24311972e+00 1.33036590e+00
-1.06944931e+00 -3.18501562e-01 3.97717625e-01 3.09116215e-01
-2.48526052e-01 -5.53463735e-02 -6.24644339e-01 6.31332576e-01
1.26662207e+00 -8.63859534e-01 6.36530697e-01 6.57013535e-01
6.84083462e-01 -4.18855220e-01 -9.16425586e-01 8.04996729e-01
-4.38826710e-01 -8.19667578e-01 -5.58322221e-02 -2.58477747e-01
6.42812490e-01 -6.75852120e-01 -3.48335415e-01 5.69921732e-01
-1.30381122e-01 -5.11290669e-01 5.55644810e-01 6.52963042e-01
6.53852701e-01 -1.50216436e+00 7.18930542e-01 2.10962489e-01
-1.53128970e+00 -5.96996188e-01 -8.91942233e-02 2.84004539e-01
-8.04164484e-02 3.69816599e-03 -8.07911932e-01 6.12187922e-01
4.27506298e-01 3.05404693e-01 -3.08271646e-01 5.81338763e-01
8.66145417e-02 3.94604117e-01 -1.57855079e-01 -9.70708191e-01
5.44107221e-02 -8.24412167e-01 5.54157138e-01 7.18451262e-01
3.65704566e-01 1.18509075e-02 -5.44278137e-02 8.57079387e-01
3.58519018e-01 3.75453651e-01 -2.78206110e-01 6.26804009e-02
8.60761523e-01 1.55529046e+00 -8.06640625e-01 -2.56368160e-01
3.70811634e-02 3.01941037e-01 -3.72971147e-01 1.90552279e-01
-1.26161981e+00 -8.48086894e-01 3.98878396e-01 2.03820795e-01
3.13524872e-01 1.07252717e-01 -4.37053815e-02 -1.18471138e-01
-4.09203738e-01 -5.91337502e-01 1.68634340e-01 -7.99850225e-01
-8.82336318e-01 2.83751637e-01 5.66584885e-01 -1.73608685e+00
-6.70696318e-01 -3.82778704e-01 -6.01429224e-01 6.25031352e-01
-1.71895897e+00 -3.86186898e-01 1.68019477e-02 6.41166747e-01
5.72797775e-01 -3.91929299e-01 3.70767295e-01 2.26942822e-01
-8.72662604e-01 5.95591128e-01 6.13068044e-01 -3.57069016e-01
8.11227560e-02 -9.83336985e-01 -7.61310697e-01 5.43696582e-01
-1.08781683e+00 -3.81243140e-01 9.09537017e-01 -3.02203387e-01
-2.03461266e+00 -9.37881052e-01 3.33271384e-01 3.05487752e-01
7.67340899e-01 1.77376643e-01 -5.03705025e-01 4.10246737e-02
6.34358943e-01 -3.94431472e-01 -5.91440387e-02 -9.46279049e-01
5.85876942e-01 -8.23020816e-01 -1.44266081e+00 6.29497290e-01
1.69127524e-01 -2.24203438e-01 -3.55870396e-01 -7.29756951e-02
4.51638252e-01 -2.04492602e-02 -1.38714910e+00 6.85008526e-01
4.62241530e-01 -3.76032472e-01 6.76933706e-01 -7.71287456e-02
-2.88735162e-02 -6.89723849e-01 1.22948185e-01 -1.70053327e+00
-3.89061451e-01 -6.84420288e-01 -4.69708815e-02 1.28412998e+00
2.68197894e-01 -7.42734730e-01 3.94446075e-01 1.36493519e-01
7.69665018e-02 -9.73403215e-01 -1.14344883e+00 -9.54456627e-01
2.93830752e-01 6.93930918e-03 7.83056438e-01 6.40061855e-01
2.57966340e-01 3.12060148e-01 -1.11100788e-03 1.93773091e-01
5.73166549e-01 -2.26681247e-01 3.58637810e-01 -6.67970538e-01
-1.51705816e-02 -4.83477890e-01 -2.41154164e-01 3.62347513e-02
1.39932588e-01 -7.72316933e-01 1.33731002e-02 -1.62708127e+00
-1.04969278e-01 -3.89311969e-01 -5.64832211e-01 3.52286279e-01
1.56467378e-01 -3.55242461e-01 1.40381664e-01 4.52753417e-02
-8.62534121e-02 8.97842169e-01 1.33018291e+00 -3.06099474e-01
-5.89038968e-01 -2.16784179e-02 -1.94604304e-02 4.60167259e-01
1.16055012e+00 -3.68083775e-01 -8.48966897e-01 3.74362737e-01
1.70898363e-01 5.68291783e-01 -3.51743726e-03 -1.36977005e+00
1.51427120e-01 -4.62298870e-01 3.58932972e-01 -8.18306386e-01
-2.67211199e-01 -1.21955752e+00 6.03522122e-01 1.13443577e+00
-1.63916335e-01 5.37460983e-01 4.02199805e-01 4.53887850e-01
2.43253056e-02 -1.94807604e-01 9.83349025e-01 4.83058095e-01
-8.85196686e-01 -1.54339969e-01 -9.26136017e-01 -2.81819463e-01
1.60214102e+00 -5.01995146e-01 -2.78628115e-02 -1.12743922e-01
-2.73951530e-01 5.32255113e-01 -1.64501086e-01 3.18850130e-01
2.36662209e-01 -1.55333185e+00 -3.56185824e-01 2.32979909e-01
-1.69736415e-01 -7.49423385e-01 2.36138195e-01 8.13132942e-01
-4.62288588e-01 3.66328597e-01 -7.77049243e-01 -5.52400887e-01
-8.42926323e-01 4.64394629e-01 7.48675942e-01 -4.05251086e-01
-2.61402309e-01 -2.10737243e-01 -5.29194593e-01 -2.60312736e-01
-2.67835379e-01 -2.47769713e-01 -6.51933014e-01 2.73339272e-01
-2.57816538e-02 9.78080571e-01 1.74034253e-01 -3.80365133e-01
-2.69509405e-01 9.70079362e-01 5.13827503e-01 -1.93737298e-02
1.10769522e+00 -1.99233949e-01 -7.56555889e-03 2.93957293e-01
1.08841074e+00 -4.73126352e-01 -1.43076789e+00 4.13335741e-01
-8.61839950e-02 1.63898915e-01 2.49616072e-01 -6.88021541e-01
-1.18743598e+00 2.07610443e-01 1.12651622e+00 2.43229046e-01
1.53634655e+00 -7.11412489e-01 4.06079233e-01 3.55793208e-01
5.31154513e-01 -2.08279872e+00 -1.40228653e-02 2.82247990e-01
7.53635466e-01 -4.57897574e-01 1.50310755e-01 -1.91722319e-01
-6.42995775e-01 1.34399068e+00 7.34403431e-01 -1.59459203e-01
7.76327610e-01 6.78146660e-01 -4.57101405e-01 3.34533423e-01
-7.16147423e-01 2.47678041e-01 -2.97022122e-03 4.34060037e-01
-1.89057350e-01 3.06279492e-02 -1.07119083e+00 4.33332741e-01
5.69062196e-02 2.10620519e-02 7.16959476e-01 8.65131140e-01
-5.10981023e-01 -1.03469133e+00 -5.79744518e-01 1.49547935e-01
1.36822835e-01 6.01064503e-01 4.66030329e-01 1.28884006e+00
1.85863420e-01 1.03647673e+00 2.28124991e-01 -5.28997123e-01
6.10537171e-01 5.32486625e-02 5.35607159e-01 1.74345449e-01
-4.93253708e-01 -5.79469018e-02 -9.95083153e-03 -4.77789015e-01
-2.58557558e-01 -4.77070570e-01 -1.84486580e+00 -2.86981195e-01
-3.02696407e-01 4.97852355e-01 7.82609880e-01 9.41824615e-01
1.48958236e-01 9.04448152e-01 1.50153077e+00 -4.71413881e-01
-7.23902464e-01 -5.07441998e-01 -7.59055018e-01 -9.47548971e-02
-7.51293451e-02 -9.87650394e-01 -6.22324944e-01 -1.81077629e-01]
|
[5.544878005981445, 2.3183176517486572]
|
04f60fab-84d8-4489-a42b-20074c64cb55
|
tstarbots-defeating-the-cheating-level
|
1809.07193
| null |
http://arxiv.org/abs/1809.07193v3
|
http://arxiv.org/pdf/1809.07193v3.pdf
|
TStarBots: Defeating the Cheating Level Builtin AI in StarCraft II in the Full Game
|
Starcraft II (SC2) is widely considered as the most challenging Real Time
Strategy (RTS) game. The underlying challenges include a large observation
space, a huge (continuous and infinite) action space, partial observations,
simultaneous move for all players, and long horizon delayed rewards for local
decisions. To push the frontier of AI research, Deepmind and Blizzard jointly
developed the StarCraft II Learning Environment (SC2LE) as a testbench of
complex decision making systems. SC2LE provides a few mini games such as
MoveToBeacon, CollectMineralShards, and DefeatRoaches, where some AI agents
have achieved the performance level of human professional players. However, for
full games, the current AI agents are still far from achieving human
professional level performance. To bridge this gap, we present two full game AI
agents in this paper - the AI agent TStarBot1 is based on deep reinforcement
learning over a flat action structure, and the AI agent TStarBot2 is based on
hard-coded rules over a hierarchical action structure. Both TStarBot1 and
TStarBot2 are able to defeat the built-in AI agents from level 1 to level 10 in
a full game (1v1 Zerg-vs-Zerg game on the AbyssalReef map), noting that level
8, level 9, and level 10 are cheating agents with unfair advantages such as
full vision on the whole map and resource harvest boosting. To the best of our
knowledge, this is the first public work to investigate AI agents that can
defeat the built-in AI in the StarCraft II full game.
|
['Bo Li', 'Yongsheng Liu', 'Peng Sun', 'Xinghai Sun', 'Ji Liu', 'Jiechao Xiong', 'Han Liu', 'Yang Zheng', 'Tong Zhang', 'Qing Wang', 'Lei Han']
|
2018-09-19
| null | null | null | null |
['real-time-strategy-games']
|
['playing-games']
|
[-3.85368675e-01 4.94325012e-01 -3.47990021e-02 2.42621034e-01
-5.98444879e-01 -6.32725835e-01 5.18832982e-01 -4.03057396e-01
-7.40802348e-01 9.57543969e-01 -3.36924404e-01 -3.79652798e-01
-5.00187099e-01 -7.17722178e-01 -4.28906620e-01 -6.82145417e-01
-2.93113142e-01 7.86593199e-01 6.77474022e-01 -1.17562938e+00
2.78029352e-01 2.45772153e-01 -1.29456091e+00 4.89793010e-02
7.28610933e-01 9.85172272e-01 2.73858428e-01 8.94073606e-01
5.50165057e-01 1.56289423e+00 -6.70765638e-01 -6.22366369e-01
8.04104388e-01 -4.06892061e-01 -8.54757547e-01 -3.38030219e-01
-1.61997154e-01 -5.97932398e-01 -4.88278687e-01 9.59755957e-01
7.09128976e-01 2.73652583e-01 4.11134288e-02 -1.69998360e+00
-1.87607091e-02 1.11318231e+00 -4.36933935e-01 1.43232197e-01
-5.48014790e-03 6.89830244e-01 1.24014878e+00 -1.76940605e-01
5.97911298e-01 1.10135329e+00 2.89869606e-01 6.13918424e-01
-4.86988664e-01 -6.62466228e-01 9.01858285e-02 5.02417445e-01
-1.02790248e+00 -1.53356299e-01 4.67201889e-01 -2.01147959e-01
1.13954115e+00 2.36482278e-01 1.03793979e+00 1.10441947e+00
4.20785010e-01 1.22094870e+00 1.20134592e+00 -2.75233220e-02
6.99496210e-01 -3.78770143e-01 -3.73492032e-01 6.43423438e-01
2.41677072e-02 8.62697840e-01 -4.04139876e-01 1.13208346e-01
8.85293961e-01 -1.32534966e-01 2.68556952e-01 -4.94176894e-01
-9.89593983e-01 1.04509711e+00 5.81468761e-01 1.32369250e-01
-7.06068814e-01 5.65768778e-01 1.28330544e-01 7.24495590e-01
-2.21747831e-01 1.16357040e+00 -2.56431967e-01 -7.76674032e-01
-6.54404342e-01 8.33791137e-01 7.80832767e-01 5.75777054e-01
4.79333907e-01 4.27496523e-01 -1.63570464e-01 2.77425587e-01
-1.36730507e-01 4.03887123e-01 3.14691037e-01 -1.55985236e+00
2.65576005e-01 4.50808227e-01 5.35237670e-01 -8.22783172e-01
-7.79495358e-01 -5.83065450e-01 -4.61784810e-01 1.06808960e+00
3.20650071e-01 -6.52092695e-01 -5.89841843e-01 1.61481059e+00
1.98540911e-01 -1.12901524e-01 3.05820674e-01 1.12858319e+00
5.84202766e-01 4.74134654e-01 -4.69506204e-01 1.12068199e-01
1.06985450e+00 -1.43611753e+00 -3.60184133e-01 -4.76013005e-01
7.13402987e-01 -1.71777546e-01 7.42456913e-01 9.24914360e-01
-1.54390430e+00 -2.34105825e-01 -1.06895137e+00 4.82168674e-01
-2.61796981e-01 -3.11212301e-01 9.54984307e-01 4.89044607e-01
-1.10573971e+00 6.08809114e-01 -1.11340904e+00 -7.15654856e-03
4.15417939e-01 5.51651478e-01 -2.56506473e-01 5.42474836e-02
-1.45335543e+00 1.17977858e+00 5.38580298e-01 -8.40945095e-02
-1.60013485e+00 -2.25019455e-01 -5.01021445e-01 1.63069293e-01
1.25470614e+00 -2.85129428e-01 1.65277958e+00 -8.52919400e-01
-1.94683838e+00 7.92202055e-01 1.10956669e+00 -9.26539838e-01
7.99559653e-01 -2.46758491e-01 -2.23036438e-01 -9.09336060e-02
1.65796623e-01 7.23481894e-01 5.02730966e-01 -7.81338453e-01
-1.20518768e+00 -2.04758674e-01 7.55376101e-01 8.75295699e-01
1.09130859e-01 7.12077990e-02 4.36514840e-02 -3.27334106e-01
-5.64723492e-01 -1.09127223e+00 -5.54133952e-01 -7.30904758e-01
-2.32034802e-01 -3.44016701e-01 2.50568599e-01 -1.12778828e-01
1.03274071e+00 -1.91732919e+00 3.24098825e-01 -1.13797225e-01
4.47108626e-01 4.45625424e-01 -2.06230283e-01 5.84901810e-01
3.94525707e-01 -1.97810426e-01 2.71084309e-01 2.36769944e-01
3.80396724e-01 4.88751322e-01 -1.43157646e-01 2.68381476e-01
-3.29731047e-01 1.22707689e+00 -1.26976597e+00 -6.67170510e-02
-8.17857161e-02 -5.62433064e-01 -7.54440963e-01 1.88042626e-01
-2.93913007e-01 3.62093411e-02 -6.09882891e-01 6.35865808e-01
3.90124321e-01 1.36164725e-01 9.92028639e-02 6.57726765e-01
-5.09455442e-01 1.27697319e-01 -1.14276779e+00 1.51357353e+00
5.35963252e-02 1.93664640e-01 5.82862139e-01 -1.06974411e+00
6.92168891e-01 1.01256549e-01 6.75056636e-01 -9.42104220e-01
3.57779801e-01 3.68798345e-01 6.85483634e-01 -1.54246345e-01
8.21666181e-01 -2.39106134e-01 -5.65640390e-01 4.19914037e-01
2.33334210e-02 -6.42124832e-01 3.73812944e-01 2.88076460e-01
1.62721515e+00 9.02706310e-02 2.24721655e-01 -2.03468353e-01
3.97699401e-02 4.44241643e-01 7.31910348e-01 1.35547483e+00
-7.00428128e-01 8.52471516e-02 8.70294034e-01 -6.68411314e-01
-7.43405402e-01 -9.25201535e-01 7.07023740e-01 1.48348510e+00
4.69556838e-01 -3.42562646e-01 -5.96714914e-01 -6.65851295e-01
-1.30850792e-01 7.99753308e-01 -5.37566304e-01 -3.63757044e-01
-4.90784913e-01 -4.09352362e-01 8.54658723e-01 3.65737289e-01
7.96556532e-01 -1.63436842e+00 -1.54099047e+00 2.83251762e-01
-9.95669663e-02 -8.28670144e-01 -1.25897527e-01 4.34046000e-01
-1.44909620e-01 -1.10948253e+00 -4.37109739e-01 -4.28538620e-01
-1.86328173e-01 2.31024325e-01 9.36919689e-01 -7.05229351e-04
-2.13814214e-01 1.82977706e-01 -6.12272084e-01 -6.58057034e-01
-1.51330858e-01 8.07780400e-02 3.64869207e-01 -6.51651859e-01
-1.08110299e-02 -3.86772722e-01 -4.75521117e-01 5.95480859e-01
-5.16810715e-01 2.06847578e-01 7.19463289e-01 1.02200139e+00
4.77032457e-03 5.39833844e-01 2.97920376e-01 -3.22258055e-01
6.87305331e-01 -3.51319492e-01 -9.70200479e-01 7.25520924e-02
-3.35306853e-01 -1.85387313e-01 5.56457818e-01 -2.25948110e-01
-8.72570097e-01 -2.33588651e-01 3.60839851e-02 -3.94737691e-01
2.20985457e-01 2.11200163e-01 3.65375280e-02 -8.25409368e-02
9.33677018e-01 1.21073835e-01 4.03066427e-02 7.40232766e-02
3.21358472e-01 3.62958163e-01 6.38087153e-01 -6.31022334e-01
7.40755677e-01 2.60121077e-01 5.00387959e-02 -3.64796072e-01
-7.22330809e-01 -7.06046522e-02 -1.77872837e-01 -3.57381761e-01
8.41685712e-01 -8.38954926e-01 -1.29477286e+00 8.82612109e-01
-6.02886438e-01 -1.04378355e+00 -6.48882806e-01 4.08702165e-01
-9.77507651e-01 -5.05952016e-02 -5.35992503e-01 -9.24359024e-01
6.25923797e-02 -1.21828127e+00 3.99717867e-01 5.23935139e-01
1.62775621e-01 -5.77154815e-01 2.03530997e-01 6.41066432e-01
4.69579905e-01 1.65277988e-01 1.73311457e-01 -7.85502493e-01
-5.83045423e-01 -4.94831353e-02 3.04801226e-01 9.73094404e-02
-2.10644782e-01 -4.28219289e-01 -2.77410477e-01 -5.60742438e-01
-5.88918477e-02 -1.08129168e+00 4.96764719e-01 4.23091888e-01
5.38836658e-01 -2.78849602e-01 -8.06066319e-02 5.68689942e-01
8.57603908e-01 7.56114841e-01 7.05684900e-01 9.37432945e-01
2.41938710e-01 3.58020931e-01 1.13252640e+00 7.19781101e-01
4.87314194e-01 7.55339861e-01 1.20529187e+00 -2.58990806e-02
1.87104434e-01 -3.82662147e-01 8.20781827e-01 2.15749368e-01
-4.55287606e-01 -4.28238332e-01 -9.02978361e-01 4.72758740e-01
-2.21366644e+00 -1.30127048e+00 1.22670636e-01 1.97932708e+00
5.90933621e-01 5.75874686e-01 5.15806019e-01 -1.67833060e-01
3.66830796e-01 1.00733161e-01 -8.14415038e-01 -6.13315225e-01
-2.16844842e-01 9.13279429e-02 7.89155066e-01 4.91540074e-01
-9.90250409e-01 1.75359762e+00 6.21136379e+00 1.23445714e+00
-7.72607625e-01 -7.36560225e-02 3.77971023e-01 -6.07819557e-01
1.40646905e-01 3.29703659e-01 -6.51362538e-01 2.89213270e-01
5.73059797e-01 -1.92575991e-01 9.85747755e-01 1.03752685e+00
9.20958593e-02 -2.48539716e-01 -6.19851172e-01 9.43785727e-01
-2.61418641e-01 -1.34747195e+00 -6.24226511e-01 4.42041099e-01
8.26515853e-01 3.72558355e-01 1.49938360e-01 9.87450600e-01
1.73020494e+00 -1.42137611e+00 7.84954011e-01 1.74733222e-01
4.56569582e-01 -1.09146035e+00 8.89465511e-01 8.06761444e-01
-6.33240819e-01 -7.16941595e-01 -4.52672988e-01 -7.14614928e-01
1.86198473e-01 -1.83493301e-01 -7.00593591e-01 6.06905103e-01
8.54613423e-01 3.03204238e-01 -8.13283324e-02 7.67944336e-01
-5.63914001e-01 4.08123195e-01 -3.23247284e-01 -2.59157419e-01
1.04560840e+00 -2.37748832e-01 7.83763170e-01 3.41056764e-01
2.08393291e-01 6.58045828e-01 7.23649800e-01 5.28869331e-01
2.84396142e-01 -4.24135178e-01 -4.61885095e-01 -3.37016463e-01
2.69238085e-01 1.39243603e+00 -6.30129933e-01 -1.05310306e-01
1.16364017e-01 8.17167819e-01 3.93501312e-01 -9.34647173e-02
-8.46678019e-01 -4.02302116e-01 8.22915137e-01 -4.89487918e-03
3.66888881e-01 -8.88757631e-02 -7.41683692e-02 -9.93789732e-01
-5.32286882e-01 -1.32475746e+00 5.38155854e-01 -8.65803003e-01
-9.41790938e-01 6.15704179e-01 -1.24649391e-01 -9.57408786e-01
-5.00017881e-01 -6.24519706e-01 -7.49868989e-01 1.77099019e-01
-1.14890087e+00 -1.00651824e+00 -6.63753822e-02 8.25104058e-01
4.50914383e-01 -9.28497016e-01 4.56197947e-01 -3.83479327e-01
-4.47137177e-01 4.97150809e-01 1.39974669e-01 -9.24308226e-02
2.18280807e-01 -1.31831110e+00 3.35747361e-01 6.70522213e-01
-4.99628037e-02 -1.01594016e-01 7.13184953e-01 -5.60830772e-01
-1.31759357e+00 -3.30574274e-01 -5.04566915e-02 -3.85885626e-01
8.91801894e-01 -2.14130849e-01 -1.72001153e-01 4.63160992e-01
2.74083287e-01 -2.19578072e-01 1.91502362e-01 -4.54917178e-02
2.80828744e-01 1.02122381e-01 -9.17524040e-01 9.40554559e-01
1.15832627e+00 4.94867228e-02 -7.44615257e-01 1.95423141e-01
5.55653751e-01 -7.69810438e-01 -3.85464281e-01 1.72822133e-01
4.16115820e-01 -1.43129158e+00 7.77106345e-01 -8.40742528e-01
6.04957402e-01 -5.09571247e-02 -1.78032473e-01 -1.69987059e+00
-6.33913755e-01 -1.22821808e+00 5.74979298e-02 4.73056167e-01
-7.04413326e-03 -4.35589284e-01 1.14229560e+00 3.01527917e-01
-3.11123163e-01 -6.87143564e-01 -1.06928277e+00 -1.07874560e+00
2.70947129e-01 -2.24575356e-01 3.97528201e-01 7.52914310e-01
4.05682296e-01 2.14179650e-01 -9.39924657e-01 -2.15139046e-01
4.59850699e-01 4.35621217e-02 1.11890888e+00 -1.00408864e+00
-8.11372757e-01 -7.19898820e-01 -3.62067193e-01 -1.00096869e+00
-1.49332851e-01 -6.75896466e-01 1.93128988e-01 -1.45875549e+00
-7.58474618e-02 -4.56904620e-01 -3.76515597e-01 8.91275406e-01
1.50285944e-01 -1.79285333e-01 6.84309483e-01 1.08280078e-01
-1.04230297e+00 6.18812740e-01 1.48490024e+00 -4.31210874e-03
-2.36232698e-01 2.45434433e-01 -9.70509171e-01 8.16546381e-01
8.87548387e-01 -3.15299690e-01 -5.73934972e-01 -1.40918955e-01
4.46864784e-01 4.82549548e-01 2.84675628e-01 -1.14253461e+00
4.43688661e-01 -9.40134168e-01 -3.41575176e-01 -3.09990674e-01
4.71739888e-01 -5.35363019e-01 -1.06148586e-01 1.07281876e+00
-2.10038573e-01 7.78765008e-02 3.02424021e-02 1.52103856e-01
-4.29015644e-02 -1.58989310e-01 6.71475053e-01 -5.46454310e-01
-1.03575182e+00 2.41464823e-01 -8.14502239e-01 4.06195670e-01
1.43640900e+00 -2.64865160e-01 -4.93079513e-01 -8.59251857e-01
-5.28826892e-01 9.52606022e-01 4.07559007e-01 3.06350470e-01
4.27669227e-01 -1.05329335e+00 -9.41828191e-01 -1.65543810e-01
-2.46632069e-01 2.74770800e-02 4.69976574e-01 7.30504274e-01
-6.99657559e-01 2.92326927e-01 -1.02114344e+00 -1.95587967e-02
-1.03101313e+00 5.05367160e-01 5.99911094e-01 -1.03349352e+00
-6.08572423e-01 1.16258180e+00 4.57223505e-01 -7.43905604e-01
6.48681447e-02 3.07113141e-01 -8.63170996e-02 -1.80006295e-01
3.54569316e-01 5.67690313e-01 -2.46921778e-01 -1.68078020e-01
-3.74334365e-01 -1.09249607e-01 -8.64001811e-02 -4.90607768e-01
1.54179037e+00 2.75357455e-01 3.38853449e-01 -3.85810472e-02
5.87375350e-02 -1.16855152e-01 -1.69852209e+00 4.35801856e-02
-2.31170237e-01 -3.38457346e-01 1.70036018e-01 -1.21814156e+00
-9.60752547e-01 7.26404667e-01 2.85283267e-01 3.59167308e-01
8.76733780e-01 -2.65701205e-01 8.19455624e-01 7.18153477e-01
7.99099028e-01 -1.69289589e+00 5.85650384e-01 9.91696298e-01
8.52918208e-01 -1.02642310e+00 -1.16537251e-01 3.97991866e-01
-1.41369569e+00 7.28379726e-01 1.09236169e+00 -4.00338113e-01
3.57529186e-02 3.77632141e-01 4.84273769e-02 -5.53261459e-01
-1.18202710e+00 -5.14822245e-01 -4.53603536e-01 7.35200047e-01
-5.25392294e-01 4.27536249e-01 -6.41556904e-02 8.31049204e-01
-8.09566379e-01 -6.13796413e-02 7.59491146e-01 1.10739803e+00
-8.16965044e-01 -9.35978234e-01 -3.35202783e-01 1.29530236e-01
-2.59602457e-01 1.20947093e-01 -6.36590779e-01 9.78373349e-01
2.33060688e-01 1.02227712e+00 -8.18893313e-02 -6.20241940e-01
3.56828898e-01 -5.80526888e-01 4.04997647e-01 -2.63635129e-01
-8.74093235e-01 -1.60395294e-01 9.13772509e-02 -8.10773730e-01
1.46926433e-01 -4.42111433e-01 -1.32965672e+00 -7.09140241e-01
-1.34732276e-01 3.28583628e-01 3.32048684e-01 9.07476187e-01
3.34491059e-02 5.18430591e-01 7.58562803e-01 -7.00052559e-01
-1.11958778e+00 -6.59546256e-01 -9.46024179e-01 4.87821735e-02
1.07423244e-02 -9.30170417e-01 -1.72337472e-01 -9.33792531e-01]
|
[3.54858660697937, 1.5538793802261353]
|
4a89c6e5-27fc-45ad-8c66-9141b87cb176
|
3dteethseg-22-3d-teeth-scan-segmentation-and
|
2305.18277
| null |
https://arxiv.org/abs/2305.18277v1
|
https://arxiv.org/pdf/2305.18277v1.pdf
|
3DTeethSeg'22: 3D Teeth Scan Segmentation and Labeling Challenge
|
Teeth localization, segmentation, and labeling from intra-oral 3D scans are essential tasks in modern dentistry to enhance dental diagnostics, treatment planning, and population-based studies on oral health. However, developing automated algorithms for teeth analysis presents significant challenges due to variations in dental anatomy, imaging protocols, and limited availability of publicly accessible data. To address these challenges, the 3DTeethSeg'22 challenge was organized in conjunction with the International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) in 2022, with a call for algorithms tackling teeth localization, segmentation, and labeling from intraoral 3D scans. A dataset comprising a total of 1800 scans from 900 patients was prepared, and each tooth was individually annotated by a human-machine hybrid algorithm. A total of 6 algorithms were evaluated on this dataset. In this study, we present the evaluation results of the 3DTeethSeg'22 challenge. The 3DTeethSeg'22 challenge code can be accessed at: https://github.com/abenhamadou/3DTeethSeg22_challenge
|
['Edouard Ladroit', 'Cyril Trosset', 'Hugo Setbon', 'Aurélien Thollot', 'Julien Strippoli', 'Shankeeth Vinayahalingam', 'Steven Kempers', 'Niels van Nistelrooij', 'Byungsun Choi', 'Jae-Hwan Han', 'Wan Kim', 'Hong-Gi Ahn', 'Tae-Hoon Yong', 'Bulat Ibragimov', 'Tudor Dascalu', 'Yuanfeng Zhou', 'Zhiming Cui', 'Guangshun Wei', 'Shaojie Zhuang', 'Juan Carlos Prieto', 'Lucia Cevidanes', 'Mathieu Leclercq', 'Yeong-Gil Shin', 'Minyoung Chung', 'Minkyung Lee', 'Minchang Kim', 'Hoyeon Lim', 'Edmond Boyer', 'Sergi Pujades', 'Ahmed Rekik', 'Oussama Smaoui', 'Achraf Ben-Hamadou']
|
2023-05-29
| null | null | null | null |
['anatomy']
|
['miscellaneous']
|
[ 1.55768663e-01 7.42645919e-01 -4.23373014e-01 -4.42393482e-01
-1.47838736e+00 -8.39385465e-02 1.24545179e-01 4.66240257e-01
-2.58394659e-01 1.96545303e-01 1.62078083e-01 -4.20964330e-01
3.39728631e-02 -4.45767671e-01 -3.63561451e-01 -7.21468508e-01
-1.75803900e-01 9.22267258e-01 2.31685609e-01 2.25479558e-01
2.79722661e-01 5.78621805e-01 -1.78883374e+00 2.24490136e-01
5.37610531e-01 7.06104875e-01 4.41904724e-01 7.40557194e-01
-2.47216970e-01 1.28047049e-01 -1.94481745e-01 -1.79415911e-01
2.68350877e-02 -8.98267701e-02 -8.22414637e-01 1.24978870e-01
7.94444621e-01 -3.40513051e-01 1.21136807e-01 1.02657640e+00
9.63675499e-01 -5.06796837e-01 6.42238438e-01 -8.62639010e-01
-2.74725378e-01 2.97688246e-01 -5.12962222e-01 3.96901697e-01
3.39785814e-01 8.49312246e-02 5.14367938e-01 -7.72725344e-01
9.64306593e-01 1.38487208e+00 5.84619701e-01 7.71256685e-01
-8.95798028e-01 -4.19693857e-01 -2.17711747e-01 3.23792458e-01
-1.05214965e+00 -6.72440946e-01 4.93188679e-01 -5.23548841e-01
5.47814548e-01 1.14900626e-01 7.65974343e-01 5.81907868e-01
6.74416870e-02 1.26083362e+00 1.21022630e+00 -8.18529248e-01
5.17222285e-01 -1.36153758e-01 5.48497081e-01 7.88645625e-01
2.14468732e-01 1.06926605e-01 -1.62964780e-02 -2.98386157e-01
5.57165444e-01 -2.51086593e-01 -8.40072110e-02 -6.47221506e-02
-8.03726256e-01 9.26825047e-01 4.90878969e-01 1.99841708e-01
-4.48847353e-01 -3.22712988e-01 4.79113430e-01 -1.01221524e-01
5.02776921e-01 -2.37478167e-01 -2.67838657e-01 2.88832277e-01
-3.88746828e-01 -3.24568041e-02 5.61040878e-01 4.81667489e-01
4.37065423e-01 -6.38564527e-01 3.92806858e-01 1.50001192e+00
8.40322673e-01 5.50102651e-01 7.56874025e-01 -1.21998656e+00
-1.53550863e-01 4.27913427e-01 -6.39835954e-01 -3.88543695e-01
-6.45629346e-01 2.65043020e-01 -4.73289251e-01 3.34956676e-01
7.22005248e-01 3.66974324e-01 -1.71218443e+00 1.11447871e+00
1.14980030e+00 -2.50878841e-01 -3.22265685e-01 4.47411060e-01
1.12212825e+00 1.41934082e-01 1.42638832e-01 -2.14944050e-01
1.73335445e+00 -7.67427444e-01 -6.61726236e-01 -2.49976501e-01
9.86380577e-01 -1.13184333e+00 9.94226396e-01 3.12001824e-01
-1.11730206e+00 -1.48022935e-01 -6.37636721e-01 -4.38943654e-01
-3.64980370e-01 -1.54465497e-01 4.80355233e-01 7.70506859e-01
-1.14137506e+00 -8.45388919e-02 -1.40735292e+00 -5.31118274e-01
9.66289282e-01 4.17909265e-01 -4.60052013e-01 -6.11536622e-01
-3.97604138e-01 1.03536415e+00 2.40792707e-02 1.72226578e-01
-3.25355798e-01 -5.50636113e-01 -8.19941521e-01 -8.62998605e-01
6.60568923e-02 -4.16511238e-01 1.77161860e+00 4.55495603e-02
-1.56353998e+00 1.74334717e+00 -3.02428454e-01 -3.95653807e-02
4.59296227e-01 3.69887054e-02 -1.88613996e-01 2.02312812e-01
7.24226534e-01 8.82915139e-01 -3.42953391e-02 -1.24693775e+00
-8.18795860e-01 -1.21957660e+00 -5.18803537e-01 -2.14208752e-01
3.21660340e-01 -4.77096438e-02 -9.58864212e-01 -3.34329516e-01
8.62652540e-01 -1.35287714e+00 -6.02521896e-01 3.50457191e-01
-7.27437198e-01 -5.92253983e-01 6.22330904e-01 -6.49560452e-01
5.49378514e-01 -2.31273937e+00 -4.98755187e-01 2.57855684e-01
9.07659605e-02 2.37417459e-01 8.26162770e-02 -5.57842925e-02
1.73857257e-01 5.54028861e-02 -3.42867732e-01 -7.39826679e-01
-3.17926884e-01 4.77576047e-01 5.66550195e-01 6.72304153e-01
-6.88059568e-01 5.06718159e-01 -8.85536671e-01 -1.04119205e+00
6.66238427e-01 6.81178272e-01 -4.91722047e-01 -2.31554002e-01
-2.60590672e-01 4.98129010e-01 -6.70045078e-01 1.37331855e+00
6.93466783e-01 -3.08628857e-01 1.79269955e-01 -2.89180160e-01
2.29752641e-02 2.54523873e-01 -8.48481178e-01 1.62882209e+00
-5.03826916e-01 2.09653601e-01 5.31217158e-01 -6.74409032e-01
3.33844185e-01 4.90490198e-01 9.63384867e-01 -9.73332524e-01
3.39310676e-01 8.94841433e-01 -2.23867461e-01 -7.81333983e-01
-2.30497897e-01 -6.08836822e-02 3.42604905e-01 6.34273469e-01
-1.79984778e-01 -5.19977272e-01 3.59758347e-01 -4.48260717e-02
8.02265823e-01 -5.74200451e-02 3.87410522e-01 -1.68297574e-01
2.18430042e-01 7.36091807e-02 3.99122626e-01 3.60884011e-01
-8.27992797e-01 7.19453633e-01 -4.55931854e-03 -5.13752162e-01
-6.00741923e-01 -1.10513234e+00 -8.84311616e-01 5.85027158e-01
-1.86883464e-01 5.53576509e-03 -9.79145527e-01 -5.71461916e-01
-5.55518121e-02 3.54041100e-01 -8.38418603e-01 3.42150599e-01
-8.78769457e-01 -8.32364261e-01 1.21769309e-01 7.96529576e-02
5.24998903e-02 -1.25831926e+00 -7.07502961e-01 4.10450011e-01
-2.20789045e-01 -1.03728044e+00 -5.86984277e-01 2.22519055e-01
-1.16555190e+00 -1.58440447e+00 -9.71360743e-01 -1.43224871e+00
8.36069584e-01 7.14940205e-02 8.42551112e-01 3.10248733e-01
-9.42561150e-01 3.88444811e-01 -6.87746257e-02 -9.54649687e-01
-6.92568898e-01 5.42344488e-02 -2.45081469e-01 -6.42542481e-01
6.60558939e-01 -4.05960768e-01 -9.21222389e-01 1.62415534e-01
-8.13134551e-01 -3.11622679e-01 5.82686543e-01 5.02584040e-01
1.25633371e+00 -5.26575863e-01 4.26319212e-01 -9.86133993e-01
2.88003027e-01 -3.13301623e-01 -5.19510984e-01 2.22007424e-01
-6.88187599e-01 -1.58688024e-01 -3.07551563e-01 -6.09057434e-02
-8.24725270e-01 5.47154784e-01 -7.18873918e-01 2.86357820e-01
-4.00607407e-01 2.37667933e-01 5.28111048e-02 1.58140808e-01
4.98079956e-01 -1.62289634e-01 5.60121000e-01 -8.06421697e-01
3.71489733e-01 1.07299387e+00 7.94217527e-01 -3.72120619e-01
4.06315029e-02 8.70700002e-01 -1.66821644e-01 -1.04180062e+00
-9.84652758e-01 -6.31188333e-01 -6.48904383e-01 -1.89831898e-01
1.01145208e+00 -6.55127645e-01 -5.20091951e-01 8.72686386e-01
-9.23129320e-01 -2.22759917e-01 -2.20810011e-01 6.54761612e-01
-3.03717822e-01 4.23794389e-01 -8.54785740e-01 -5.40352881e-01
-5.76877415e-01 -1.75920892e+00 1.28022456e+00 3.48307222e-01
-2.61495799e-01 -9.53316033e-01 3.95683348e-01 1.03303206e+00
-3.71844247e-02 3.68772358e-01 1.04324937e+00 -2.95421958e-01
-1.19599685e-01 -2.83514321e-01 8.40298384e-02 2.99214512e-01
6.38010859e-01 1.01974726e-01 -9.96851802e-01 1.62960142e-01
3.89197953e-02 -3.27521861e-01 7.74107218e-01 1.24244654e+00
9.56539094e-01 3.91025878e-02 -7.24182785e-01 2.79968351e-01
9.25335526e-01 5.56047142e-01 6.54602408e-01 7.41597950e-01
3.73333961e-01 9.90027487e-01 3.18990648e-01 8.11060425e-03
9.38826859e-01 5.62942088e-01 6.37803078e-01 -3.05949897e-02
-7.06753910e-01 1.67557657e-01 -1.76187977e-01 8.82015109e-01
1.00369155e-01 1.95747033e-01 -1.20167315e+00 8.62182975e-01
-1.05115175e+00 -3.71452361e-01 -1.92893475e-01 1.96398842e+00
1.01014173e+00 -5.58228381e-02 1.03548869e-01 2.95216918e-01
8.64630997e-01 -3.36661100e-01 -8.01361084e-01 -5.63377440e-02
3.18483740e-01 6.03938818e-01 2.75114000e-01 7.71870255e-01
-8.95466208e-01 6.20308280e-01 6.24127054e+00 5.84229350e-01
-1.19048262e+00 2.29608580e-01 9.45816457e-01 -5.97185157e-02
-1.00262120e-01 -4.52854872e-01 -9.46859956e-01 3.32835227e-01
7.82881498e-01 2.68209040e-01 -4.02574122e-01 8.66101563e-01
2.33997777e-01 -4.19208765e-01 -8.43986154e-01 6.85322285e-01
-1.38052758e-02 -1.12389994e+00 -3.66951138e-01 6.41048670e-01
5.20601213e-01 5.47423422e-01 1.21488675e-01 -1.29440680e-01
2.70693630e-01 -7.34539926e-01 3.38754773e-01 -6.72946870e-02
6.21045649e-01 -2.52236754e-01 3.74176174e-01 -2.09606767e-01
-8.62081289e-01 2.75195092e-01 2.59030968e-01 7.24059522e-01
5.51915944e-01 9.03786302e-01 -1.23692811e+00 -8.48824009e-02
7.76047945e-01 2.67310262e-01 -2.08246857e-01 1.35577095e+00
-1.74358875e-01 3.30504417e-01 -7.41038620e-01 3.35713804e-01
8.37709941e-03 1.98050663e-02 2.61108875e-01 6.22618139e-01
3.86420459e-01 3.59974921e-01 8.45346972e-02 -4.80191316e-03
-1.26608610e-01 2.73383141e-01 -1.11987025e-01 5.25142729e-01
6.29114211e-01 1.01593459e+00 -1.16414857e+00 -3.52927089e-01
-2.32438207e-01 3.63395691e-01 -2.35246584e-01 -2.28418633e-01
-1.70550913e-01 3.30154628e-01 3.31178099e-01 4.69602734e-01
-6.14506789e-02 1.48456335e-01 -3.82828981e-01 -5.37244558e-01
-1.61900088e-01 -9.49158669e-01 7.01725364e-01 -5.36042452e-01
-1.20174313e+00 5.40798426e-01 8.85321125e-02 -7.64330387e-01
-2.42361590e-01 -7.54173219e-01 -3.42808217e-01 2.86843687e-01
-1.37939262e+00 -1.20485783e+00 -2.47485012e-01 3.99857104e-01
5.85877955e-01 2.59647787e-01 1.03580177e+00 5.53923726e-01
-5.82526207e-01 4.20041621e-01 2.39692956e-01 -1.77743047e-01
4.86822426e-01 -1.12339473e+00 1.18910067e-01 -1.20231202e-02
1.30282417e-02 3.41839522e-01 4.84061629e-01 -5.62606215e-01
-1.10086429e+00 -7.79767036e-01 8.62864494e-01 -3.41471165e-01
1.84383094e-01 2.33276919e-01 -5.82180262e-01 7.93372393e-01
-1.21372022e-01 2.61767894e-01 9.74838972e-01 -2.23791122e-01
-7.11878911e-02 3.38766724e-01 -1.63292372e+00 4.22567695e-01
6.89419985e-01 -3.75508130e-01 -2.79664159e-01 8.14540088e-01
3.80350918e-01 -8.26209664e-01 -1.18854403e+00 4.81997401e-01
7.98479378e-01 -7.13565946e-01 1.12446129e+00 2.40645006e-01
2.25448370e-01 2.05812827e-02 -3.88486564e-01 -6.69282317e-01
4.68475521e-01 -1.32467538e-01 1.67594522e-01 6.35637045e-01
2.88324416e-01 -8.59660625e-01 1.13842237e+00 3.55688214e-01
-7.80613244e-01 -9.81273055e-01 -1.50934470e+00 -3.05571556e-01
4.89042073e-01 -4.31972712e-01 2.07742974e-01 6.70458496e-01
-3.60732436e-01 -2.09018379e-01 5.40903091e-01 1.89701736e-01
1.06293952e+00 -2.90565155e-02 4.13464159e-01 -1.25926375e+00
1.33901775e-01 -5.57642519e-01 -3.95018160e-01 -7.22505987e-01
-3.41357708e-01 -8.84982347e-01 1.62753418e-01 -2.09375858e+00
1.56092584e-01 -6.05504513e-01 -1.18333802e-01 7.78700829e-01
-2.66970452e-02 5.53706110e-01 -2.74064302e-01 4.75607336e-01
2.40217552e-01 4.95342128e-02 1.64508379e+00 -2.82820791e-01
-2.52878845e-01 3.08202595e-01 -4.94000524e-01 7.44107962e-01
7.10407078e-01 -5.92437446e-01 -2.26418480e-01 -3.78640294e-01
-6.52681053e-01 -5.56811206e-02 2.85911918e-01 -7.71145940e-01
2.18738452e-01 1.61157355e-01 -2.03458965e-01 -9.61130738e-01
5.24540007e-01 -3.60990822e-01 3.97092514e-02 9.81076896e-01
1.80919886e-01 -4.08518702e-01 1.17477156e-01 1.07031316e-01
-3.59287798e-01 -1.43328369e-01 1.08795631e+00 -5.22242904e-01
-4.83324915e-01 3.61921132e-01 -2.51287103e-01 -6.56142756e-02
1.03467786e+00 -4.90770936e-01 -3.21445614e-01 8.36225152e-02
-1.30198789e+00 1.70816049e-01 2.36000553e-01 1.56334177e-01
4.62723851e-01 -9.48895097e-01 -7.49659061e-01 8.01691562e-02
2.03029677e-01 6.63009942e-01 5.99569678e-01 1.00271177e+00
-8.21253002e-01 4.47434276e-01 -1.71382558e-02 -1.03396976e+00
-1.43562353e+00 -6.10521100e-02 4.53700840e-01 -2.14239508e-01
-7.19328463e-01 9.24772680e-01 1.86950825e-02 -8.30620348e-01
1.71837449e-01 -5.54698646e-01 -1.89422861e-01 2.02426299e-01
4.62785274e-01 3.82084996e-01 4.44020003e-01 -9.12639141e-01
-5.14336944e-01 1.00261152e+00 -6.46360815e-01 7.17771947e-02
1.48803031e+00 -1.95334747e-01 -1.00470372e-02 2.21039504e-02
1.47263038e+00 -1.90028995e-01 -5.39721191e-01 -2.30675653e-01
1.73818693e-01 -1.76192090e-01 5.60402453e-01 -7.48215795e-01
-1.39953220e+00 6.56362593e-01 1.48414576e+00 4.96522896e-02
8.79813969e-01 7.07868338e-01 1.20823395e+00 -2.91191578e-01
4.50495511e-01 -9.11146581e-01 -8.20345953e-02 -1.00075481e-02
7.24396050e-01 -1.66441035e+00 8.00088868e-02 -7.41443694e-01
-2.62352428e-03 1.00196874e+00 2.52058893e-01 4.18632835e-01
1.32268512e+00 2.33816996e-01 8.68381083e-01 -7.35811651e-01
5.29671647e-02 -1.54827591e-02 -2.36133356e-02 7.46046782e-01
6.92889571e-01 1.19182982e-01 -4.32020068e-01 -2.30167359e-01
-4.50253248e-01 2.04584330e-01 2.31939286e-01 1.41263497e+00
-3.91729534e-01 -1.55208349e+00 -5.57013273e-01 3.74533534e-01
-6.86155856e-01 1.96117282e-01 1.47666410e-01 8.48812521e-01
1.70124948e-01 1.03306770e+00 -9.69334841e-02 3.25521767e-01
3.92190367e-01 -3.79900396e-01 5.49341321e-01 -7.86252558e-01
1.22416094e-02 3.87537628e-01 5.61268926e-02 -5.67335844e-01
-5.37258744e-01 -1.10430467e+00 -1.65515864e+00 2.79807262e-02
-2.74642974e-01 -1.22250922e-01 1.30323923e+00 8.55688155e-01
2.06042245e-01 3.47353131e-01 5.32877684e-01 -9.05410171e-01
-3.18998992e-01 -9.62402165e-01 -5.02112865e-01 1.13338917e-01
3.53493541e-01 -6.07083321e-01 -6.18628561e-01 2.44263336e-01]
|
[13.740005493164062, -2.2246017456054688]
|
b1a20084-279e-41aa-b27b-aa7e33773690
|
temporal-context-mining-for-learned-video
|
2111.13850
| null |
https://arxiv.org/abs/2111.13850v2
|
https://arxiv.org/pdf/2111.13850v2.pdf
|
Temporal Context Mining for Learned Video Compression
|
We address end-to-end learned video compression with a special focus on better learning and utilizing temporal contexts. For temporal context mining, we propose to store not only the previously reconstructed frames, but also the propagated features into the generalized decoded picture buffer. From the stored propagated features, we propose to learn multi-scale temporal contexts, and re-fill the learned temporal contexts into the modules of our compression scheme, including the contextual encoder-decoder, the frame generator, and the temporal context encoder. Our scheme discards the parallelization-unfriendly auto-regressive entropy model to pursue a more practical decoding time. We compare our scheme with x264 and x265 (representing industrial software for H.264 and H.265, respectively) as well as the official reference software for H.264, H.265, and H.266 (JM, HM, and VTM, respectively). When intra period is 32 and oriented to PSNR, our scheme outperforms H.265--HM by 14.4% bit rate saving; when oriented to MS-SSIM, our scheme outperforms H.266--VTM by 21.1% bit rate saving.
|
['Yan Lu', 'Dong Liu', 'Li Li', 'Bin Li', 'Jiahao Li', 'Xihua Sheng']
|
2021-11-27
| null | null | null | null |
['ms-ssim']
|
['computer-vision']
|
[ 3.16531658e-01 -1.87472492e-01 -2.57982492e-01 -3.39943349e-01
-7.67923892e-01 -2.82449811e-03 6.11950085e-02 -8.18674453e-03
-4.23691124e-01 6.56341553e-01 2.91750580e-01 -3.61136258e-01
4.87310477e-02 -5.17446399e-01 -6.15141451e-01 -7.66869426e-01
-7.81150043e-01 -6.27855897e-01 2.48436734e-01 2.18759239e-01
3.37532431e-01 -5.57214655e-02 -1.38738203e+00 6.85578167e-01
3.08807671e-01 1.49378085e+00 7.89323449e-01 1.16430581e+00
5.32162964e-01 1.45182014e+00 -3.29718322e-01 -2.53604472e-01
1.62168205e-01 -5.29041469e-01 -5.70252001e-01 2.65590101e-01
4.94699292e-02 -8.45978320e-01 -8.50713551e-01 8.12559247e-01
3.05474341e-01 3.84109803e-02 1.86931804e-01 -8.07542026e-01
-2.04368129e-01 6.41021490e-01 -6.54233932e-01 3.82734746e-01
2.45892450e-01 3.10811818e-01 8.52310240e-01 -8.15650940e-01
5.96840560e-01 1.05005789e+00 3.63359272e-01 4.53230470e-01
-7.28773117e-01 -6.61330938e-01 3.23678851e-02 8.19406152e-01
-1.43252158e+00 -6.46411002e-01 6.69028878e-01 -1.49176121e-01
1.16457605e+00 2.66657978e-01 6.91630661e-01 9.83137250e-01
7.50626624e-01 7.19537377e-01 5.83946824e-01 -3.50782424e-01
4.15428936e-01 -5.07667959e-01 -1.80432752e-01 6.02943480e-01
-2.53486782e-01 3.64856899e-01 -6.74842060e-01 6.91136569e-02
7.38622129e-01 -5.08921556e-02 -4.35940057e-01 2.46992409e-01
-1.41787708e+00 4.84448940e-01 4.35682647e-02 7.04103485e-02
-5.74034810e-01 5.26845098e-01 7.23540246e-01 6.42836511e-01
2.07896456e-01 -3.24455321e-01 -7.64718652e-01 -6.38097942e-01
-1.42145288e+00 -1.89958259e-01 5.07340014e-01 1.32168972e+00
7.96990156e-01 2.88451791e-01 -1.59890071e-01 6.56120300e-01
4.85459656e-01 3.34760159e-01 6.14755630e-01 -1.50423324e+00
7.43749261e-01 1.62270367e-02 -2.14637578e-01 -8.76932204e-01
9.97828990e-02 -2.27572724e-01 -1.15081584e+00 -1.57028660e-01
-2.99439162e-01 -2.56443977e-01 -6.72596514e-01 1.47105134e+00
-1.76765010e-01 7.20719755e-01 2.05179349e-01 8.08364749e-01
2.95101106e-01 1.07285452e+00 5.22913709e-02 -9.07418072e-01
1.10369766e+00 -9.72970665e-01 -7.63321698e-01 -1.16445765e-01
5.76149583e-01 -8.13235939e-01 6.34307206e-01 4.84118909e-01
-1.50547385e+00 -8.75039995e-01 -1.35833538e+00 -1.14206895e-01
4.70861495e-01 2.78328329e-01 -5.02858078e-04 4.81668264e-02
-1.17622125e+00 8.88042212e-01 -8.83275867e-01 4.65922505e-02
2.76724756e-01 4.52058911e-01 2.48420238e-02 -8.91482383e-02
-1.09824932e+00 3.46662909e-01 7.21472740e-01 -3.32654625e-01
-1.27672362e+00 -4.81464088e-01 -1.01471448e+00 2.39452377e-01
3.74721467e-01 -4.84636843e-01 1.32176566e+00 -8.24911237e-01
-1.45361376e+00 3.64065796e-01 -4.77938205e-01 -8.06550741e-01
1.68191850e-01 -2.89590776e-01 -7.18770623e-01 5.24975717e-01
-2.94944316e-01 6.48074329e-01 1.06734109e+00 -1.06180632e+00
-1.09369636e+00 1.42141283e-02 2.70625819e-02 4.26714830e-02
-2.95419842e-01 -7.41813704e-02 -8.71979713e-01 -1.01120925e+00
8.45319405e-02 -8.43546212e-01 -2.52959311e-01 -3.12970802e-02
-1.57656610e-01 3.76008630e-01 1.25580704e+00 -1.22983575e+00
1.89557588e+00 -2.54239869e+00 1.26217335e-01 5.47070056e-02
2.46428326e-01 2.09378302e-01 -1.51706357e-02 6.03218041e-02
1.89005416e-02 -1.01664662e-03 -2.68426418e-01 -4.10185933e-01
-3.27074170e-01 4.18085605e-01 -2.51856923e-01 3.38633865e-01
-7.96691254e-02 5.24439633e-01 -8.42712283e-01 -7.36170352e-01
3.56309444e-01 6.05098307e-01 -8.22778165e-01 3.01720619e-01
-1.63575262e-01 3.52032810e-01 -2.75754660e-01 6.20137036e-01
4.95945364e-01 -2.69901872e-01 5.64586759e-01 -4.61791158e-01
-1.43729180e-01 3.88292700e-01 -9.56408799e-01 1.90316248e+00
-6.49917901e-01 8.95590067e-01 5.06495940e-04 -7.23678708e-01
7.38820016e-01 6.93702281e-01 6.91805184e-01 -8.02306235e-01
1.45239467e-02 1.51675358e-01 -2.90094018e-01 -6.46298349e-01
8.78361523e-01 1.94523305e-01 2.07514375e-01 -1.72033738e-02
1.53545380e-01 3.95887941e-01 3.52234334e-01 1.70684651e-01
1.21173286e+00 1.69574350e-01 5.25730491e-01 3.91457492e-04
7.90849447e-01 -7.15594828e-01 8.72610271e-01 1.76693305e-01
-1.50747046e-01 6.10458493e-01 3.77647281e-01 -3.30649763e-01
-1.34544933e+00 -9.19097126e-01 2.16205165e-01 9.95625675e-01
2.10902005e-01 -9.74497259e-01 -5.56647182e-01 -3.88741165e-01
-4.40791756e-01 8.15056205e-01 -1.57070085e-01 -2.88167059e-01
-1.01347470e+00 -2.08777145e-01 3.22929651e-01 3.80957216e-01
8.33163977e-01 -9.08814073e-01 -1.00623155e+00 5.52721560e-01
-4.41504747e-01 -1.25857294e+00 -8.24192822e-01 1.05455488e-01
-1.26444066e+00 -8.06035757e-01 -6.03792727e-01 -7.80045211e-01
3.11064832e-02 1.56943709e-01 1.10148799e+00 2.68068947e-02
2.09496692e-02 2.57695884e-01 -4.97301221e-01 2.89313525e-01
-5.77147841e-01 -3.25895429e-01 -2.27097765e-01 -5.87594174e-02
7.26263970e-02 -8.41503799e-01 -8.96273017e-01 5.19575179e-01
-8.77780914e-01 3.32878381e-01 5.93487680e-01 7.85977185e-01
8.41775775e-01 2.06225723e-01 1.42821535e-01 -3.05785626e-01
-8.26604143e-02 -6.12697542e-01 -4.54894245e-01 2.28332937e-01
-7.90496409e-01 -9.64328796e-02 5.91999054e-01 -3.59081179e-01
-1.00786030e+00 -8.73357132e-02 -1.04329713e-01 -7.92745352e-01
2.08327428e-01 4.24313277e-01 -2.16093406e-01 2.03191966e-01
1.98878258e-01 4.79648650e-01 -3.69818360e-01 -3.12657624e-01
1.32243350e-01 6.41345680e-01 7.98917413e-01 -3.14557672e-01
5.05908847e-01 1.41057834e-01 -1.20171331e-01 -7.53707767e-01
-1.31799862e-01 -3.34483922e-01 -2.69088745e-01 -2.75328517e-01
8.63598406e-01 -1.31707060e+00 -5.68637013e-01 1.43386364e-01
-1.12541020e+00 -3.39808196e-01 -3.41987610e-01 7.93675423e-01
-1.06369495e+00 7.31522143e-01 -1.15834236e+00 -6.05910897e-01
-3.91253173e-01 -1.32264817e+00 1.01707125e+00 -1.10951893e-01
-8.97310898e-02 -7.08984613e-01 -3.99056941e-01 1.39615446e-01
4.05617625e-01 1.85918771e-02 8.56181979e-01 1.94709972e-02
-8.55080605e-01 3.16124856e-01 -1.65321782e-01 7.07113564e-01
9.14729666e-03 5.60672954e-02 -8.14564884e-01 -5.35629869e-01
4.44287568e-01 -4.55772802e-02 9.40862596e-01 5.09505510e-01
1.54914784e+00 -7.19167709e-01 5.43598412e-03 1.03427720e+00
1.50122023e+00 8.39603007e-01 9.34014559e-01 2.55159497e-01
4.08954620e-01 1.87762290e-01 7.83499002e-01 1.13140357e+00
3.86248738e-01 6.25001073e-01 4.93203223e-01 3.51908296e-01
-2.70630270e-01 -1.47440910e-01 9.72783089e-01 1.63621449e+00
-3.52101654e-01 -3.91485393e-01 -5.87882638e-01 2.73586750e-01
-1.67806351e+00 -1.20896387e+00 3.56829524e-01 2.10474396e+00
9.58384216e-01 2.52996773e-01 -3.24275762e-01 5.17798603e-01
6.11678064e-01 5.31181693e-01 -4.63426590e-01 -3.12216341e-01
-6.16684146e-02 5.35777658e-02 4.65018779e-01 4.40635353e-01
-9.98972058e-01 4.43645597e-01 5.51632547e+00 1.04128671e+00
-1.13911009e+00 6.12535551e-02 9.58346188e-01 -3.48109752e-01
-7.33432397e-02 2.85394549e-01 -4.05842960e-01 7.72290826e-01
1.50751269e+00 -2.96612680e-01 6.78807080e-01 8.31298709e-01
5.48800766e-01 -3.56066301e-02 -1.35672653e+00 1.41347289e+00
-7.77902305e-02 -1.45669687e+00 2.03811028e-03 1.01115562e-01
6.27979934e-01 -1.53422222e-01 7.57666603e-02 3.89615744e-01
-3.80345732e-01 -6.94095373e-01 1.00398457e+00 4.58898395e-01
1.17308939e+00 -7.14646041e-01 5.99684358e-01 1.60717830e-01
-1.69378579e+00 -4.08636928e-01 -3.81878495e-01 1.31216556e-01
4.25021440e-01 7.09034681e-01 -3.81579667e-01 6.19068086e-01
7.83429980e-01 1.22530174e+00 -3.51276129e-01 6.69240832e-01
1.11508273e-01 7.86930025e-01 9.37492177e-02 6.21554613e-01
1.30187780e-01 5.95762730e-02 4.54466045e-01 1.42832899e+00
7.63934970e-01 2.50493377e-01 1.35725304e-01 1.56468898e-01
-9.39194337e-02 -1.24298364e-01 -1.72956407e-01 2.66814798e-01
2.63797790e-01 6.77202761e-01 -9.51574966e-02 -8.09387565e-01
-6.44557714e-01 1.13668048e+00 -3.01099420e-01 5.12257397e-01
-1.15126777e+00 -1.09016642e-01 6.52022421e-01 -8.17465931e-02
6.27197504e-01 -5.21551728e-01 -1.42937619e-02 -1.26255310e+00
1.50880367e-01 -1.10472620e+00 3.99599522e-01 -7.16595292e-01
-5.71300745e-01 4.36754107e-01 7.79934600e-02 -1.67253470e+00
-5.79258204e-01 -1.88238844e-01 -2.88110614e-01 5.53724051e-01
-1.59082544e+00 -6.19903505e-01 -2.44350165e-01 8.50235820e-01
1.19776988e+00 -3.48824114e-01 3.65487754e-01 6.34042621e-01
-2.57874638e-01 5.68278015e-01 7.24608526e-02 -6.52171895e-02
5.80850840e-01 -6.04405940e-01 2.86961675e-01 9.46444213e-01
-1.38998553e-01 2.73493439e-01 5.80003619e-01 -4.92519587e-01
-1.63418686e+00 -1.45220339e+00 9.15976465e-01 5.97736001e-01
4.05528069e-01 8.36514980e-02 -8.50131333e-01 6.07031643e-01
4.02727246e-01 2.29440182e-01 5.09854257e-01 -7.13134170e-01
-1.58973277e-01 -4.20608789e-01 -1.07462943e+00 5.93895555e-01
9.47928309e-01 -6.67904675e-01 -2.94615567e-01 -5.38583174e-02
1.08999181e+00 -4.48711783e-01 -1.15166247e+00 5.09539843e-01
5.54685652e-01 -1.24598920e+00 8.76163602e-01 5.90105169e-02
8.46277356e-01 -1.88438386e-01 -9.47245002e-01 -5.36787570e-01
-3.99072558e-01 -8.92230213e-01 -9.48341727e-01 9.95282114e-01
1.75838336e-01 -4.19335579e-03 7.02321708e-01 4.05185334e-02
-2.43636861e-01 -8.56532514e-01 -1.16607153e+00 -7.50837266e-01
-4.18628693e-01 -7.61460364e-01 5.37746966e-01 5.74174106e-01
-9.15478691e-02 -8.15440044e-02 -8.36390913e-01 1.81454659e-01
6.25211179e-01 7.45261693e-03 2.48837605e-01 -2.36537606e-01
-8.82340491e-01 -2.34679297e-01 -5.33153057e-01 -1.62501776e+00
-1.66984245e-01 -4.74234760e-01 -2.56930832e-02 -8.28018606e-01
2.30238155e-01 -2.09278669e-02 -7.13678718e-01 1.79079071e-01
1.08929142e-01 -8.73773471e-02 3.61325532e-01 3.21480811e-01
-7.30829239e-01 6.38146222e-01 1.10696900e+00 -2.95642287e-01
-1.79429781e-02 -1.38208568e-01 -1.43507347e-01 6.77828670e-01
1.01294112e+00 -1.83470964e-01 -7.16000199e-01 -5.65014899e-01
-1.99685827e-01 7.99059570e-01 2.78282344e-01 -1.29715312e+00
2.09930256e-01 -2.04512611e-01 3.38243216e-01 -6.68048680e-01
4.28289086e-01 -9.17316973e-01 2.44126186e-01 5.75628817e-01
-4.13045555e-01 4.32832837e-01 -1.54538415e-02 5.94380796e-01
-5.40744543e-01 6.22794172e-03 7.73036838e-01 2.05442905e-02
-1.11207688e+00 4.54790950e-01 -4.46731120e-01 -1.03535309e-01
1.02559841e+00 -2.62276232e-01 -8.82092640e-02 -5.11539400e-01
-9.08029139e-01 -1.28637189e-02 5.03303647e-01 2.40467489e-01
1.24300277e+00 -1.28078508e+00 -7.17493415e-01 2.63898283e-01
-1.94471970e-01 -2.51994878e-01 4.89978760e-01 6.77124560e-01
-6.16626918e-01 2.32987016e-01 4.02357392e-02 -6.31495059e-01
-1.46058774e+00 7.30381012e-01 -1.18302755e-01 -3.79432738e-01
-6.63259089e-01 3.42137694e-01 1.15383029e-01 8.28159213e-01
4.49519634e-01 -6.83759689e-01 2.17771642e-02 -3.65734488e-01
7.18692124e-01 6.21467590e-01 -1.19914480e-01 -5.89564860e-01
-1.34404317e-01 4.22374636e-01 1.32106408e-01 -4.42930385e-02
1.19896722e+00 -6.59341812e-01 -8.42755660e-03 4.03599352e-01
1.66402483e+00 -1.96898535e-01 -1.57884324e+00 -1.25998229e-01
5.08030802e-02 -5.03606141e-01 7.11058527e-02 -5.30716300e-01
-1.48284400e+00 7.09860981e-01 8.17408741e-01 -2.51343399e-01
1.87784600e+00 -3.07149202e-01 1.14243233e+00 1.84281126e-01
6.60363376e-01 -9.62045908e-01 7.36716464e-02 5.99815726e-01
8.19776535e-01 -9.63353217e-01 3.27213079e-01 -2.23836839e-01
-5.10007739e-01 1.30256724e+00 2.74630815e-01 1.25710085e-01
8.56913805e-01 4.86322016e-01 -3.03025901e-01 4.01256502e-01
-1.45860422e+00 3.40269983e-01 1.05875924e-01 2.77367860e-01
5.20336151e-01 -5.25771044e-02 -2.66829789e-01 2.73672640e-01
4.31202650e-02 2.81774521e-01 5.08926690e-01 1.12052941e+00
-4.30076718e-01 -1.01075327e+00 -2.71150231e-01 3.37459117e-01
-5.22099853e-01 -2.99950510e-01 6.35065496e-01 4.46815103e-01
3.61283809e-01 1.02608645e+00 6.10773712e-02 -9.63908911e-01
6.08481746e-03 -1.78474143e-01 2.47274384e-01 -7.28280693e-02
-3.45193803e-01 4.54430014e-01 5.64569123e-02 -9.26457822e-01
-5.54395318e-01 -6.66496396e-01 -1.32709146e+00 -5.01721323e-01
1.95947871e-01 -1.66657537e-01 4.81286198e-01 6.39798820e-01
3.71702075e-01 6.94234967e-01 1.07971632e+00 -8.50476325e-01
-3.47695231e-01 -7.40152776e-01 -5.75458288e-01 2.97024846e-01
7.31873274e-01 -9.92213469e-03 -3.64425659e-01 8.55517805e-01]
|
[11.332518577575684, -1.6449073553085327]
|
65520799-d12b-4443-a06c-46814587afb9
|
humor-generation-and-detection-in-code-mixed
| null | null |
https://aclanthology.org/2021.ranlp-srw.1
|
https://aclanthology.org/2021.ranlp-srw.1.pdf
|
Humor Generation and Detection in Code-Mixed Hindi-English
|
Computational humor generation is one of the hardest tasks in natural language generation, especially in code-mixed languages. Existing research has shown that humor generation in English is a promising avenue. However, studies have shown that bilingual speakers often appreciate humor more in code-mixed languages with unexpected transitions and clever word play. In this study, we propose several methods for generating and detecting humor in code-mixed Hindi-English. Of the experimented approaches, an Attention Based Bi-Directional LSTM with converting parts of text on a word2vec embedding gives the best results by generating 74.8% good jokes and IndicBERT used for detecting humor in code-mixed Hindi-English outperforms other humor detection methods with an accuracy of 96.98%.
|
['Rhythm Narula', 'Kaustubh Agarwal']
| null | null | null | null |
ranlp-2021-9
|
['humor-detection']
|
['natural-language-processing']
|
[-7.75177479e-01 3.46479230e-02 -2.57838294e-02 1.68471217e-01
-3.20490330e-01 -3.06462377e-01 5.73157847e-01 -1.32480502e-01
8.35398491e-03 8.39749694e-01 9.36705172e-01 -4.74298060e-01
5.21049380e-01 -6.67186022e-01 -2.48627111e-01 -2.22977161e-01
3.52145523e-01 7.90278390e-02 -3.59460622e-01 -9.03839707e-01
7.47383237e-01 5.83001077e-02 -1.07688415e+00 5.72545946e-01
1.04548597e+00 -3.09904981e-02 1.76802576e-01 8.33876967e-01
-3.45576584e-01 2.26839638e+00 -8.89159560e-01 -7.94417739e-01
-1.52584657e-01 -1.03754580e+00 -1.01953614e+00 -1.22004323e-01
3.22736144e-01 -3.12913328e-01 -6.35647357e-01 1.20785224e+00
7.47699499e-01 1.25369979e-02 4.97507244e-01 -1.08150065e+00
-1.23827767e+00 1.31637323e+00 -4.67635185e-01 1.47842482e-01
7.14943051e-01 3.00329924e-01 9.25853908e-01 -1.33079648e+00
8.29954803e-01 1.21497595e+00 7.71120012e-01 7.01351881e-01
-8.79404128e-01 -7.13895917e-01 -1.08257520e+00 5.73617876e-01
-1.07837927e+00 -1.65972769e-01 1.19300890e+00 -9.25248325e-01
1.29967499e+00 -3.56831332e-03 4.40587550e-01 1.10584176e+00
6.66636825e-01 1.05863655e+00 1.14354527e+00 -5.30416787e-01
-1.82144359e-01 6.66795373e-01 6.02598861e-02 1.01125300e+00
2.93405131e-02 -2.76260436e-01 -8.27590466e-01 -1.56032413e-01
3.31890374e-01 -2.73481309e-01 -1.95218042e-01 5.48309207e-01
-1.05985594e+00 1.46825314e+00 5.11036098e-01 7.69687653e-01
-2.29511097e-01 3.51470448e-02 9.56320524e-01 5.31599641e-01
9.64738354e-02 8.56942236e-01 3.24184120e-01 -4.65872079e-01
-1.18106246e+00 3.74226302e-01 1.02285600e+00 1.17826927e+00
3.55371416e-01 7.40917921e-01 -2.01702788e-01 1.14107275e+00
8.73049945e-02 4.96409863e-01 1.27760005e+00 -6.12710357e-01
4.56234008e-01 7.31272161e-01 5.39627969e-02 -1.49493504e+00
-2.81062812e-01 -6.10020101e-01 -8.34932327e-01 4.34921026e-01
3.45639512e-02 -6.29652143e-02 -4.39762712e-01 1.34634173e+00
-6.13210201e-01 -8.96180451e-01 -1.80995967e-02 9.34838712e-01
1.25890303e+00 7.88112938e-01 -2.14711890e-01 2.31060199e-02
1.15953684e+00 -1.52340138e+00 -8.71555567e-01 -3.68099630e-01
5.70676982e-01 -1.30266297e+00 1.52487087e+00 2.15032801e-01
-1.20416570e+00 -5.15885532e-01 -1.18632209e+00 -5.05654812e-01
-3.68409663e-01 2.84720480e-01 1.11737944e-01 5.28996229e-01
-8.33763182e-01 4.63333309e-01 -8.34003370e-03 -3.41530621e-01
1.34668872e-01 -3.64489824e-01 -2.21870899e-01 2.04637557e-01
-1.38667476e+00 1.44615746e+00 1.94981754e-01 -3.92438918e-01
-1.00446618e+00 -1.24783330e-01 -1.02011251e+00 -8.28812923e-03
-4.13707554e-01 -3.57562512e-01 1.36439991e+00 -1.12481761e+00
-1.21214223e+00 1.02681041e+00 -2.21843049e-02 -4.88930374e-01
4.96734142e-01 -1.99377269e-01 -4.11233008e-01 -1.46477222e-01
4.31610793e-01 1.28984287e-01 7.81162918e-01 -1.08677912e+00
-1.45380482e-01 2.60936171e-01 4.03657220e-02 -8.61890465e-02
-6.23515844e-01 3.40125680e-01 5.77060759e-01 -6.04735374e-01
-3.19784015e-01 -6.17251635e-01 7.74023533e-02 -7.37957954e-01
-4.37778771e-01 -2.38027364e-01 4.48851436e-01 -1.38585854e+00
1.83711135e+00 -1.80009186e+00 4.63907123e-02 -4.57492650e-01
6.24326050e-01 1.96039289e-01 -1.97585188e-02 9.67876196e-01
1.51801765e-01 1.15241788e-01 -1.56874552e-01 -2.15901136e-02
2.24179342e-01 -7.97693506e-02 -5.25960565e-01 3.23713928e-01
-3.98486070e-02 9.69729125e-01 -1.28189230e+00 -5.93617141e-01
-1.63091980e-02 2.26573080e-01 -4.55631673e-01 4.54430342e-01
3.52486134e-01 -1.49308830e-01 1.41004562e-01 5.59354186e-01
4.43206966e-01 -1.78074434e-01 5.92214195e-03 5.37737489e-01
-4.80118364e-01 6.88901246e-01 -3.94235611e-01 1.23332953e+00
-8.33854795e-01 1.43732893e+00 -4.55045104e-01 -1.28355980e-01
1.21968198e+00 3.78493309e-01 -2.60271400e-01 -6.47807837e-01
1.93071768e-01 6.75745487e-01 3.73165429e-01 -8.88860762e-01
1.03641415e+00 -6.24615371e-01 -2.81323552e-01 5.10659754e-01
1.46506757e-01 -3.69435519e-01 3.60330254e-01 5.94543338e-01
1.19321227e+00 -2.80443907e-01 8.09115231e-01 -4.49456364e-01
8.74865472e-01 1.71444446e-01 2.51194298e-01 8.03254247e-01
-3.11838984e-01 5.55898130e-01 6.84906662e-01 -5.78172565e-01
-1.49647260e+00 -6.00166380e-01 2.20677763e-01 1.09977114e+00
-5.38157523e-01 -5.17081022e-01 -6.78796709e-01 -3.48844469e-01
-1.79191351e-01 1.23770475e+00 -5.21597922e-01 -9.22631025e-02
-6.14501834e-01 -3.41385216e-01 8.40872109e-01 4.92868960e-01
4.95269865e-01 -1.71919489e+00 -7.70703852e-01 4.46390122e-01
-5.27653456e-01 -6.92782044e-01 -5.42324781e-01 1.12566680e-01
-4.26048130e-01 -8.41369152e-01 -7.15529919e-01 -1.01001215e+00
4.47574645e-01 3.64718109e-01 1.41842234e+00 5.98473489e-01
-3.69597107e-01 -3.92570645e-01 -6.41255677e-01 -1.76134914e-01
-1.23222053e+00 -3.37840579e-02 -1.88772127e-01 -7.48288751e-01
6.59219503e-01 -3.10072213e-01 -9.20135081e-02 -2.93637335e-01
-6.52832329e-01 2.25267261e-01 4.97373849e-01 1.29806197e+00
-8.76656711e-01 -3.29656482e-01 6.54545069e-01 -8.94090772e-01
1.41821480e+00 -8.44211102e-01 1.26445740e-01 -1.84008166e-01
-6.38034284e-01 -3.45789716e-02 1.30630660e+00 -2.43313625e-01
-8.84710431e-01 -3.92941296e-01 -7.41496533e-02 4.38166037e-02
1.23241715e-01 7.32538640e-01 4.94230449e-01 -1.72931906e-02
1.43432844e+00 4.53425914e-01 -1.46392271e-01 -9.84425247e-02
3.57560754e-01 1.11732697e+00 3.44899058e-01 -2.65530288e-01
8.53741348e-01 -9.08823907e-02 -5.79851329e-01 -8.99848819e-01
-5.05840421e-01 -6.25092924e-01 -5.02110541e-01 -4.99544829e-01
7.21925139e-01 -9.57545638e-01 -3.30807477e-01 7.05066442e-01
-1.65843689e+00 -2.47404560e-01 1.50990307e-01 1.50405571e-01
-4.08027947e-01 6.14893913e-01 -1.35230958e+00 -9.82357740e-01
-7.78451502e-01 -8.94088089e-01 3.18589598e-01 1.87508136e-01
-8.04559410e-01 -8.42738986e-01 3.88104051e-01 5.79234481e-01
6.89433336e-01 2.33493060e-01 1.24148548e+00 -4.08997864e-01
-5.91897964e-03 -1.69640839e-01 -2.27026358e-01 3.77528280e-01
-8.73254612e-02 -2.61884183e-02 -8.16583693e-01 -9.19376034e-03
2.92671919e-01 -7.83571959e-01 6.29713535e-01 -2.71306306e-01
2.29906812e-01 -1.05561161e+00 5.91712773e-01 2.54000966e-02
1.50047445e+00 1.77952066e-01 9.42755520e-01 5.16178370e-01
6.02475107e-01 4.18046772e-01 1.76751480e-01 8.96448016e-01
4.24340010e-01 1.76079422e-01 1.83650583e-01 1.24993272e-01
-4.31419611e-01 -5.52192152e-01 9.53161359e-01 1.59548604e+00
4.44495827e-02 1.29180193e-01 -1.32902753e+00 9.52984512e-01
-1.62586832e+00 -1.55725718e+00 -1.02417207e+00 1.75096774e+00
1.32455254e+00 -5.31893484e-02 3.55216980e-01 2.19054893e-01
4.71462816e-01 2.97231793e-01 7.79547170e-02 -1.03468812e+00
-3.03857684e-01 1.77307636e-01 1.05760381e-01 9.73123968e-01
-4.79065120e-01 1.14664781e+00 5.94685268e+00 4.43329543e-01
-1.05669916e+00 4.09370959e-01 2.06456408e-01 1.17542423e-01
-5.45681059e-01 -7.37193450e-02 -1.61833465e-01 5.80142260e-01
6.73940897e-01 -6.40784085e-01 6.54985547e-01 1.34313774e+00
4.55513567e-01 -3.51492390e-02 -5.91370702e-01 1.08263743e+00
7.89452970e-01 -9.69005167e-01 -2.04461515e-02 -3.43678266e-01
1.18587744e+00 -1.10454559e-01 -1.01799801e-01 7.15096235e-01
4.27350491e-01 -1.33549237e+00 1.01340103e+00 2.23327786e-01
3.30810755e-01 -6.92141891e-01 1.11301541e+00 4.54684168e-01
-5.34319341e-01 -2.78079629e-01 -7.27852464e-01 -8.12555492e-01
1.23984642e-01 6.39804304e-01 -1.11318600e+00 -3.66380930e-01
4.26037878e-01 5.66470623e-01 -9.72083151e-01 9.52644229e-01
-7.59539843e-01 8.07577252e-01 5.88895380e-01 -7.43661642e-01
3.62771749e-01 1.04921557e-01 6.83095098e-01 1.90204585e+00
3.68448138e-01 -3.34266484e-01 -2.49004334e-01 1.34776986e+00
-1.48195356e-01 5.54424524e-01 -9.62937653e-01 -3.14202756e-01
1.20463148e-01 1.43187869e+00 -2.13427857e-01 -4.67394888e-01
-2.85434932e-01 1.16500187e+00 4.09727454e-01 3.42669673e-02
-1.03118873e+00 -8.89021575e-01 7.12661222e-02 8.20368677e-02
-3.27977598e-01 -3.92960757e-01 -7.42703021e-01 -1.06729722e+00
-1.47807732e-01 -1.29904807e+00 -5.68034165e-02 -1.11859405e+00
-1.25153553e+00 9.34379697e-01 -6.47878587e-01 -1.16526282e+00
-3.78275543e-01 -6.76757455e-01 -1.01083624e+00 8.73322785e-01
-8.98722529e-01 -1.35166585e+00 -5.47328472e-01 3.14273328e-01
1.11447752e+00 -6.76203847e-01 5.74938238e-01 2.89921939e-01
-1.53506055e-01 4.73988086e-01 -6.95867464e-02 4.32086110e-01
6.37785494e-01 -1.47923124e+00 2.95176268e-01 1.03202176e+00
-1.16748452e-01 9.56401110e-01 1.19677401e+00 -6.93289220e-01
-9.62361634e-01 -6.21488154e-01 1.84931457e+00 -3.90880466e-01
1.09645510e+00 -1.79586381e-01 -8.84829760e-01 3.38369727e-01
1.10796082e+00 -1.03802419e+00 6.63391948e-01 -1.08472094e-01
-6.89359486e-01 4.72713262e-01 -9.51782703e-01 6.49485826e-01
5.07837474e-01 -1.00406194e+00 -8.67286921e-01 4.62831616e-01
2.65492648e-01 -1.29565567e-01 -4.00003254e-01 -3.27787250e-01
3.73533428e-01 -1.26232243e+00 2.22603634e-01 -7.48031557e-01
1.89515388e+00 -3.31775486e-01 8.24109092e-02 -1.40002811e+00
-7.88952291e-01 -7.60785997e-01 -5.26694395e-02 9.92419183e-01
3.79425228e-01 -3.10277995e-02 3.18448097e-01 2.30448231e-01
-2.49671891e-01 -2.31499493e-01 -4.36691761e-01 -5.11954248e-01
5.84897697e-01 1.31539837e-01 2.47925445e-02 1.37730968e+00
9.24275517e-01 8.01397622e-01 -1.07366133e+00 -8.25911045e-01
3.90303075e-01 3.86118814e-02 8.77806664e-01 -5.57272673e-01
-1.93379864e-01 -8.36417258e-01 -4.39391136e-01 -6.34577215e-01
4.34759170e-01 -1.17918098e+00 3.11623633e-01 -1.50916123e+00
9.09718454e-01 5.98011315e-01 2.78173089e-01 4.75417614e-01
-1.93368554e-01 2.63868660e-01 4.34347332e-01 1.28176659e-01
-1.45098180e-01 4.11037028e-01 1.06068230e+00 -3.33354771e-01
-3.05146724e-02 -4.57965881e-01 -6.93732858e-01 5.93940794e-01
1.06642222e+00 -5.71193278e-01 -5.69136143e-02 -1.25006527e-01
6.77118003e-01 1.17238410e-01 3.34122300e-01 -1.14337194e+00
1.96959019e-01 -2.93489069e-01 -3.11436616e-02 -3.32042366e-01
-3.41200590e-01 -1.05451994e-01 -1.46094307e-01 8.34219277e-01
-4.64338094e-01 2.88664252e-01 2.51938291e-02 -2.31102660e-01
-4.65164363e-01 -9.84412789e-01 1.14703250e+00 -5.75303495e-01
-4.92589116e-01 -7.16530204e-01 -1.11797118e+00 2.59264737e-01
4.69932884e-01 5.08603044e-02 -6.90614462e-01 -1.02259016e+00
2.85543948e-02 -3.35568219e-01 6.31278574e-01 7.27093577e-01
8.09901953e-01 -1.54537690e+00 -1.28381181e+00 -1.83175161e-01
1.56030566e-01 -1.15379500e+00 2.56488454e-02 8.21996927e-01
-1.11378074e+00 3.98264050e-01 -7.44864941e-01 2.56442606e-01
-1.23638248e+00 5.71689785e-01 3.13625365e-01 -1.36088893e-01
-3.64162028e-01 6.11818254e-01 -2.54580200e-01 -3.94476026e-01
-3.36432368e-01 4.38811570e-01 -3.30130696e-01 6.30140305e-02
5.46508253e-01 8.00452471e-01 -1.19686693e-01 -1.12418604e+00
-7.92245120e-02 -1.57112032e-01 3.21733057e-02 -1.49942800e-01
1.01888001e+00 2.48723015e-01 -7.88025379e-01 7.34264910e-01
1.17391872e+00 6.86489582e-01 -3.81909125e-02 2.82500416e-01
1.40396982e-01 -6.63796067e-01 -3.69666427e-01 -9.58919406e-01
-6.20899439e-01 1.16888905e+00 6.68975338e-02 2.83928871e-01
5.56409657e-01 -3.82290632e-01 1.06214345e+00 3.36217195e-01
4.55596298e-01 -1.51119280e+00 5.42662501e-01 1.24269426e+00
1.48775172e+00 -1.10020912e+00 -1.36288613e-01 4.48682278e-01
-9.90401745e-01 1.48171508e+00 9.74894881e-01 -2.06339374e-01
4.59896438e-02 -1.27674015e-02 3.64072710e-01 -2.96050251e-01
-5.99094093e-01 1.08063035e-01 1.62939906e-01 1.76219195e-01
1.57465780e+00 2.10953578e-01 -1.12645781e+00 6.49788618e-01
-8.68889987e-01 -2.40263477e-01 1.53676844e+00 4.87451911e-01
-8.22160661e-01 -6.73360050e-01 -4.05886322e-01 3.50197017e-01
-5.79115212e-01 -6.89984798e-01 -9.99889851e-01 5.55504441e-01
-6.15028106e-02 1.12329984e+00 -7.69371390e-01 -8.21133971e-01
-3.87229249e-02 4.07932341e-01 4.89260023e-03 -6.74202561e-01
-1.36563480e+00 -4.43266779e-01 1.94915503e-01 -1.36789650e-01
3.39858681e-02 -2.75321275e-01 -1.21656466e+00 -9.50029671e-01
-2.21975461e-01 2.41164640e-01 3.30197066e-01 6.47336245e-01
-9.52347964e-02 3.05037498e-01 6.53947473e-01 -3.62372249e-01
-7.31576860e-01 -1.55614424e+00 -5.50256908e-01 8.91395807e-01
2.20167965e-01 -1.67039156e-01 -6.51399612e-01 2.13687539e-01]
|
[8.886561393737793, 11.037076950073242]
|
eb591c93-0f20-42bb-8cb5-efd925f2fae1
|
realistic-pointgoal-navigation-via-auxiliary
|
2109.08677
| null |
https://arxiv.org/abs/2109.08677v1
|
https://arxiv.org/pdf/2109.08677v1.pdf
|
Realistic PointGoal Navigation via Auxiliary Losses and Information Bottleneck
|
We propose a novel architecture and training paradigm for training realistic PointGoal Navigation -- navigating to a target coordinate in an unseen environment under actuation and sensor noise without access to ground-truth localization. Specifically, we find that the primary challenge under this setting is learning localization -- when stripped of idealized localization, agents fail to stop precisely at the goal despite reliably making progress towards it. To address this we introduce a set of auxiliary losses to help the agent learn localization. Further, we explore the idea of treating the precise location of the agent as privileged information -- it is unavailable during test time, however, it is available during training time in simulation. We grant the agent restricted access to ground-truth localization readings during training via an information bottleneck. Under this setting, the agent incurs a penalty for using this privileged information, encouraging the agent to only leverage this information when it is crucial to learning. This enables the agent to first learn navigation and then learn localization instead of conflating these two objectives in training. We evaluate our proposed method both in a semi-idealized (noiseless simulation without Compass+GPS) and realistic (addition of noisy simulation) settings. Specifically, our method outperforms existing baselines on the semi-idealized setting by 18\%/21\% SPL/Success and by 15\%/20\% SPL in the realistic setting. Our improved Success and SPL metrics indicate our agent's improved ability to accurately self-localize while maintaining a strong navigation policy. Our implementation can be found at https://github.com/NicoGrande/habitat-pointnav-via-ib.
|
['Erik Wijmans', 'Dhruv Batra', 'Guillermo Grande']
|
2021-09-17
| null | null | null | null |
['pointgoal-navigation']
|
['robots']
|
[-5.15855066e-02 1.76389903e-01 -1.08657360e-01 -1.89621091e-01
-9.67031419e-01 -1.01845479e+00 3.70998204e-01 1.49997830e-01
-1.11031163e+00 1.10613847e+00 -3.34260076e-01 -5.02566576e-01
-1.84691295e-01 -8.59951675e-01 -1.41482413e+00 -8.46531451e-01
-7.81740904e-01 4.21141297e-01 4.65279957e-03 -1.94885224e-01
6.81944424e-03 3.53568822e-01 -1.11585343e+00 -7.54809320e-01
7.87656426e-01 6.94798172e-01 3.85943234e-01 8.97325456e-01
6.26790226e-01 5.55681467e-01 -5.82265079e-01 3.02893311e-01
5.68446517e-01 -1.19809859e-01 -4.20988709e-01 -4.00082767e-01
3.52959871e-01 -6.54846668e-01 -4.09036160e-01 1.06836677e+00
5.45923233e-01 3.64899188e-01 2.31947631e-01 -1.33158803e+00
2.66734790e-03 4.42420304e-01 -3.32687825e-01 1.59715652e-01
3.20021451e-01 6.04855239e-01 7.14664578e-01 -4.39085625e-02
4.39083934e-01 7.39241421e-01 7.09451199e-01 6.14791453e-01
-1.17818904e+00 -6.28262460e-01 5.85269928e-01 -4.50898677e-01
-1.18258119e+00 -4.09562737e-01 2.92936772e-01 -2.73660392e-01
7.53734350e-01 1.33836195e-02 4.94141161e-01 1.30259001e+00
2.45238259e-01 4.92172092e-01 8.06059420e-01 1.52190194e-01
7.56918788e-01 -1.66537568e-01 -1.38117850e-01 7.17340350e-01
4.63562340e-01 4.69278306e-01 -4.07331169e-01 -1.13805301e-01
8.65127325e-01 8.22344050e-03 -5.39682806e-01 -6.05859935e-01
-1.32865322e+00 5.64952433e-01 8.41703534e-01 -1.65355667e-01
-6.15655422e-01 6.93394721e-01 -4.57698405e-02 4.30207759e-01
-2.03785136e-01 6.26132905e-01 -4.91585851e-01 -5.71223140e-01
-4.78835821e-01 5.86065888e-01 8.07100177e-01 1.23027873e+00
8.06963980e-01 2.54992455e-01 3.20747316e-01 1.23782791e-01
1.23629652e-01 9.35904503e-01 1.66188210e-01 -1.29082167e+00
8.66440654e-01 2.19491482e-01 9.27300453e-01 -7.22719669e-01
-5.97932875e-01 -8.68032396e-01 -4.90102798e-01 4.89841193e-01
5.52028000e-01 -9.03214812e-01 -1.09322786e+00 2.50364780e+00
4.31439579e-01 4.36607480e-01 3.83576334e-01 1.15710318e+00
-3.06336209e-02 5.08215547e-01 -1.91215426e-02 6.51563182e-02
8.61363828e-01 -6.46569788e-01 -2.99714416e-01 -8.90994012e-01
7.60432601e-01 -1.71390712e-01 1.06284344e+00 1.50493875e-01
-7.57882476e-01 -2.15194777e-01 -1.33571589e+00 5.42327523e-01
-1.70095727e-01 -2.57066816e-01 6.20506823e-01 4.06174123e-01
-1.20746887e+00 5.93542278e-01 -1.46728492e+00 -5.46799958e-01
2.71445423e-01 7.11359799e-01 -4.94091809e-01 -1.61534935e-01
-1.02620351e+00 7.81096220e-01 3.09004206e-02 -3.20255794e-02
-1.49030864e+00 -5.69338381e-01 -1.07789195e+00 -2.08473638e-01
5.44682562e-01 -8.80380332e-01 1.44079685e+00 -4.50907260e-01
-1.34963274e+00 9.89769027e-02 -4.63348515e-02 -9.34783101e-01
7.14357078e-01 -5.79598248e-01 1.02771103e-01 -1.19237386e-01
5.36397398e-01 6.66258156e-01 4.23009217e-01 -1.60706830e+00
-9.21688259e-01 -3.74207526e-01 6.32473469e-01 6.12998426e-01
2.52265960e-01 -1.16394138e+00 -4.08494860e-01 -2.70589054e-01
2.89435118e-01 -1.24847651e+00 -5.07357240e-01 2.35303082e-02
-3.70296776e-01 3.87634456e-01 6.69233203e-01 -6.56864643e-02
6.13990247e-01 -1.97731090e+00 -1.28436282e-01 3.27249497e-01
1.41253740e-01 -1.63266525e-01 -1.60383433e-01 5.78304470e-01
4.49378252e-01 6.62655160e-02 -1.99468434e-01 -6.86013401e-01
1.76383238e-02 5.15510976e-01 -5.47421336e-01 6.26768410e-01
-3.12901318e-01 6.89662337e-01 -1.23915315e+00 1.00126363e-01
2.75852561e-01 4.92275983e-01 -6.76632226e-01 1.63683653e-01
-1.59783050e-01 9.54653144e-01 -9.24271882e-01 5.39737940e-01
5.22665322e-01 -8.99106041e-02 2.47084439e-01 3.98613870e-01
-2.56657958e-01 3.18537414e-01 -1.25347304e+00 1.90577090e+00
-7.06748784e-01 2.37139151e-01 5.52795172e-01 -6.72809064e-01
3.92242074e-01 -1.22958355e-01 3.71263891e-01 -7.87052333e-01
4.62163240e-02 1.88597918e-01 4.92939614e-02 -2.22513393e-01
3.37077707e-01 2.29933485e-01 -3.16834807e-01 3.03533763e-01
-1.68566361e-01 -1.32159114e-01 -1.76965460e-01 2.21386731e-01
1.53977919e+00 4.01601464e-01 -8.27082619e-02 -1.12393446e-01
-2.61181816e-02 2.48029172e-01 6.28207624e-01 1.41289628e+00
-2.44494051e-01 1.53569713e-01 2.44816840e-01 -2.57231981e-01
-7.66627371e-01 -1.19130397e+00 1.41917974e-01 9.40825522e-01
7.19105184e-01 -2.46063933e-01 -6.39590144e-01 -5.71575522e-01
3.24605793e-01 7.14314938e-01 -6.56979620e-01 -9.10289139e-02
-7.01969326e-01 -4.98420686e-01 6.33427382e-01 3.84678602e-01
6.53252006e-01 -4.70786422e-01 -1.02221286e+00 2.44357914e-01
2.80533498e-03 -8.04774344e-01 -3.32049698e-01 6.28443480e-01
-6.67814195e-01 -9.78193760e-01 -3.59801590e-01 -4.64343101e-01
9.24959898e-01 4.14049864e-01 6.20651901e-01 1.26090214e-01
1.82244703e-01 6.64118946e-01 -1.10145085e-01 -1.48345575e-01
1.48373678e-01 1.83190763e-01 4.80822384e-01 -4.80095744e-01
-2.66786695e-01 -7.45642662e-01 -9.47087228e-01 3.75154525e-01
-5.58770537e-01 -2.30395675e-01 5.06009519e-01 6.92427754e-01
6.24933243e-01 8.51706415e-03 3.27561140e-01 -5.05191624e-01
4.30492729e-01 -7.27844954e-01 -1.12810111e+00 -3.58431876e-01
-3.87706906e-01 2.47972891e-01 7.09922373e-01 -4.15633917e-01
-4.13179785e-01 2.02612221e-01 -2.07443442e-02 -1.92736521e-01
-1.11979179e-01 3.85743082e-01 -1.46507338e-01 -4.59460527e-01
6.84265018e-01 2.44239628e-01 -3.63727883e-02 -2.74968326e-01
7.56884590e-02 1.29127249e-01 5.84198952e-01 -9.37200546e-01
1.12253451e+00 6.19320631e-01 1.21430486e-01 -4.43640471e-01
-5.00694335e-01 -4.09916751e-02 4.12752526e-03 4.05026972e-02
3.83969456e-01 -1.11011004e+00 -1.32410264e+00 2.05314517e-01
-5.77607214e-01 -1.03098845e+00 -2.21741900e-01 5.99178314e-01
-5.58527887e-01 6.46000504e-02 -2.12453008e-01 -9.53495026e-01
8.55179951e-02 -1.26042140e+00 1.04249346e+00 5.33461928e-01
1.54144362e-01 -9.53409731e-01 1.46222010e-01 -6.38578981e-02
6.50825560e-01 4.20116752e-01 2.80723602e-01 -5.14956653e-01
-9.25881922e-01 -2.51605809e-01 3.46706480e-01 -1.81820929e-01
8.40910375e-02 -7.91115046e-01 -7.63654470e-01 -8.77106726e-01
-1.57357395e-01 -3.36676747e-01 7.45263755e-01 3.43149304e-01
6.76610827e-01 -5.01104891e-01 -7.31595516e-01 8.13556612e-01
1.45115268e+00 3.86091381e-01 4.02777120e-02 7.15745807e-01
3.32005888e-01 1.73456952e-01 8.15211892e-01 4.62859422e-01
9.02438998e-01 6.22333825e-01 1.07914352e+00 5.69158494e-02
4.15409714e-01 -6.77221298e-01 3.62588942e-01 -5.00611439e-02
1.22439511e-01 -5.70251167e-01 -8.38232756e-01 5.89589536e-01
-2.12174678e+00 -6.27062619e-01 3.79608959e-01 2.77447414e+00
5.46493649e-01 4.74090517e-01 -1.19329959e-01 -4.54036534e-01
2.84657836e-01 1.51235998e-01 -1.07840717e+00 2.34082639e-01
1.60726786e-01 -3.08853894e-01 1.30447710e+00 1.13893712e+00
-1.19691670e+00 1.03876257e+00 5.27514648e+00 1.99292198e-01
-1.28716266e+00 -4.89402115e-02 3.38262200e-01 -3.07267368e-01
3.60468216e-02 2.05312148e-01 -8.74046981e-01 5.23019373e-01
9.20669854e-01 1.24710307e-01 6.86856389e-01 1.18734908e+00
3.06996852e-01 -6.84413433e-01 -1.16320288e+00 6.01431012e-01
-3.36649239e-01 -1.00538194e+00 -5.86793840e-01 2.37255335e-01
4.99182552e-01 4.32130277e-01 2.95203775e-01 5.12408674e-01
8.29414248e-01 -8.39652598e-01 8.58947575e-01 2.65660971e-01
5.65928698e-01 -5.99689782e-01 8.56937945e-01 6.91413879e-01
-1.13383520e+00 -1.59043014e-01 -1.72890406e-02 -3.68790269e-01
2.79092252e-01 -9.03202146e-02 -9.88624990e-01 4.33564037e-01
5.68236291e-01 3.68200034e-01 -2.37276316e-01 1.13986742e+00
-5.21894038e-01 5.47481298e-01 -9.08263981e-01 4.48650122e-02
6.27771318e-01 -5.50045483e-02 7.51782179e-01 5.99022806e-01
3.20678353e-01 -1.83747206e-02 6.84670389e-01 6.62491143e-01
1.60683677e-01 -6.64686978e-01 -8.03247988e-01 3.86231840e-01
8.20604980e-01 8.49443734e-01 -4.14852321e-01 -5.84942661e-03
1.14695281e-01 8.22019517e-01 3.04219604e-01 8.13736022e-01
-1.03090584e+00 -4.45607364e-01 1.04877675e+00 1.88392624e-02
8.33563954e-02 -6.94686115e-01 -1.03454605e-01 -9.85712707e-01
1.62932292e-01 -3.84204626e-01 6.91103488e-02 -4.56174582e-01
-6.75813138e-01 5.72362542e-01 -7.99348727e-02 -1.21986532e+00
-5.70581496e-01 -2.80621886e-01 -3.66964310e-01 8.28936577e-01
-1.62382281e+00 -1.09100795e+00 -4.06959325e-01 4.61129844e-01
2.28101537e-01 1.35410845e-01 7.05344677e-01 2.22484440e-01
-2.99830735e-01 5.96293628e-01 1.26393110e-01 1.18077733e-02
6.64925337e-01 -1.45905876e+00 4.64061975e-01 9.29176211e-01
-2.86480278e-01 1.07061315e+00 1.06129169e+00 -8.51490438e-01
-1.83435023e+00 -1.19780850e+00 7.51407146e-02 -5.91961503e-01
5.26413739e-01 -5.06511271e-01 -5.57116508e-01 1.14142537e+00
-3.50034624e-01 7.72088245e-02 1.16640955e-01 -1.09484948e-01
-5.23821600e-02 -1.84100926e-01 -1.25977266e+00 8.48116457e-01
1.04010224e+00 -7.10891932e-02 -1.83414251e-01 3.15890580e-01
9.04792786e-01 -8.48414838e-01 -5.26980400e-01 3.79390895e-01
5.09958208e-01 -5.78116119e-01 8.77595723e-01 -2.87791133e-01
-4.16789472e-01 -7.74541855e-01 -3.28213304e-01 -1.40111947e+00
1.65975332e-01 -9.62030709e-01 2.05273598e-01 6.80502057e-01
4.84940439e-01 -1.01247621e+00 1.05352712e+00 5.33643782e-01
-1.63696155e-01 -5.46784520e-01 -1.28908038e+00 -7.37249553e-01
6.11194633e-02 -5.17979920e-01 4.44974095e-01 5.49337029e-01
-2.04118267e-01 2.97423620e-02 -4.36668426e-01 1.08902335e+00
8.64436030e-01 -2.37808526e-01 1.11447811e+00 -6.35158300e-01
-2.02008232e-01 2.25515738e-02 -2.76286870e-01 -1.52247000e+00
6.73473105e-02 -3.39169919e-01 4.53357279e-01 -1.55013251e+00
-4.50244069e-01 -1.02913392e+00 -8.69032741e-02 7.14900136e-01
1.16403453e-01 -1.22695744e-01 1.80880073e-02 1.76523075e-01
-9.17648256e-01 4.05183524e-01 9.94227767e-01 7.77602643e-02
-4.27679867e-01 4.61609513e-01 -5.84456384e-01 4.77081984e-01
8.96396339e-01 -4.83742774e-01 -6.68055654e-01 -7.53445089e-01
1.81522056e-01 3.17574382e-01 7.12457597e-01 -1.49276006e+00
6.69612646e-01 -3.02856535e-01 2.18750432e-01 -2.55351514e-01
7.42594361e-01 -1.04843855e+00 8.85837898e-02 7.36701131e-01
-2.43293375e-01 1.59452260e-01 3.16241086e-01 9.89803493e-01
2.98436761e-01 8.16177204e-02 5.55562615e-01 -1.42039374e-01
-7.31582463e-01 3.05278569e-01 -3.60499114e-01 9.75978747e-02
1.10448384e+00 -1.22969970e-01 -5.80080271e-01 -8.06575418e-01
-6.48184597e-01 9.22266901e-01 9.60101008e-01 1.91255242e-01
3.32347065e-01 -8.54181767e-01 -9.25699547e-02 2.77343094e-01
-2.77637634e-02 3.79131317e-01 2.66519547e-01 7.50877202e-01
-5.37110090e-01 4.08639073e-01 -2.01384258e-02 -5.66158354e-01
-6.28667533e-01 2.03923911e-01 7.59933352e-01 6.55864701e-02
-3.88401568e-01 8.20987701e-01 2.71994919e-01 -6.76769793e-01
6.02519035e-01 -5.52690983e-01 3.01909775e-01 -6.93310201e-01
3.45656633e-01 1.76271554e-02 -1.50965989e-01 -3.63422602e-01
-4.98093009e-01 5.00647366e-01 -3.23685296e-02 -4.72250819e-01
1.24634624e+00 -3.35105032e-01 3.73959899e-01 2.25603595e-01
7.15461612e-01 1.46373913e-01 -2.00715876e+00 2.80129407e-02
-2.15218157e-01 -3.77389580e-01 4.44996655e-02 -1.02074766e+00
-5.56267083e-01 4.43825096e-01 7.70457268e-01 -7.41199628e-02
5.67423820e-01 -1.78277269e-01 4.78078365e-01 9.29409862e-01
1.07763851e+00 -8.33529055e-01 -1.60738200e-01 7.50648379e-01
4.75405276e-01 -1.44186568e+00 -1.96564764e-01 1.46487907e-01
-5.57214499e-01 5.52805841e-01 9.24702287e-01 -4.08901036e-01
3.20872784e-01 4.37778383e-01 1.69304252e-01 4.51473556e-02
-6.40311241e-01 -2.86132276e-01 -4.64993715e-01 8.69302809e-01
-2.74166852e-01 8.04164335e-02 3.80623281e-01 5.08875132e-01
-4.76353139e-01 -3.73967499e-01 4.57907915e-01 1.54723287e+00
-7.48467267e-01 -7.85416126e-01 -3.26323837e-01 1.48708463e-01
-2.41346866e-01 4.58948538e-02 6.59553856e-02 1.19559038e+00
-4.61750142e-02 8.90362144e-01 -3.10989097e-03 -4.49384302e-01
3.83308649e-01 -4.38956708e-01 9.35955793e-02 -5.33476949e-01
-1.99635893e-01 -5.59374988e-02 -9.17047448e-03 -9.80805635e-01
8.00667629e-02 -4.63459402e-01 -1.53266072e+00 -3.03518236e-01
-1.81833428e-04 5.32300055e-01 8.98969769e-01 6.68838441e-01
6.55821204e-01 5.08304596e-01 4.22035426e-01 -1.11804950e+00
-7.36836016e-01 -6.39020860e-01 -1.91022679e-01 -1.68176636e-01
1.06771481e+00 -8.46888483e-01 -6.41176045e-01 -4.17059004e-01]
|
[4.583775520324707, 0.7603144645690918]
|
ba9687d1-907d-4407-93e1-093c7a9e99ec
|
visual-spatio-temporal-relation-enhanced
|
2110.15609
| null |
https://arxiv.org/abs/2110.15609v3
|
https://arxiv.org/pdf/2110.15609v3.pdf
|
BiC-Net: Learning Efficient Spatio-Temporal Relation for Text-Video Retrieval
|
The task of text-video retrieval aims to understand the correspondence between language and vision, has gained increasing attention in recent years. Previous studies either adopt off-the-shelf 2D/3D-CNN and then use average/max pooling to directly capture spatial features with aggregated temporal information as global video embeddings, or introduce graph-based models and expert knowledge to learn local spatial-temporal relations. However, the existing methods have two limitations: 1) The global video representations learn video temporal information in a simple average/max pooling manner and do not fully explore the temporal information between every two frames. 2) The graph-based local video representations are handcrafted, it depends heavily on expert knowledge and empirical feedback, which may not be able to effectively mine the higher-level fine-grained visual relations. These limitations result in their inability to distinguish videos with the same visual components but with different relations. To solve this problem, we propose a novel cross-modal retrieval framework, Bi-Branch Complementary Network (BiC-Net), which modifies transformer architecture to effectively bridge text-video modalities in a complementary manner via combining local spatial-temporal relation and global temporal information. Specifically, local video representations are encoded using multiple transformer blocks and additional residual blocks to learn spatio-temporal relation features, calling the module a Spatio-Temporal Residual transformer (SRT). Meanwhile, Global video representations are encoded using a multi-layer transformer block to learn global temporal features. Finally, we align the spatio-temporal relation and global temporal features with the text feature on two embedding spaces for cross-modal text-video retrieval.
|
['Hao Chen', 'Chuhao Shi', 'Yawen Zeng', 'Guangyi Xiao', 'Jingjing Chen', 'Ning Han']
|
2021-10-29
| null | null | null | null |
['video-similarity']
|
['computer-vision']
|
[-1.64278626e-01 -6.70450449e-01 -5.22668123e-01 -1.81686208e-01
-5.76772511e-01 -5.12598634e-01 8.36708546e-01 3.09309419e-02
-3.34614515e-01 2.12158635e-01 3.44573557e-01 1.87933445e-02
-4.84639913e-01 -6.98496103e-01 -4.42047209e-01 -5.60892463e-01
-1.75325453e-01 -5.11059947e-02 4.52217132e-01 -1.74548224e-01
7.68190399e-02 4.26102310e-01 -1.45292115e+00 5.17184675e-01
5.46567678e-01 1.32003343e+00 1.44128174e-01 6.27496421e-01
-3.92633110e-01 9.05517936e-01 -1.70799404e-01 -4.60532829e-02
2.39048794e-01 -3.93334985e-01 -6.27016008e-01 -1.05529949e-02
3.98414433e-01 -4.28238332e-01 -1.11548364e+00 9.28754210e-01
3.29211593e-01 3.49851280e-01 6.53851748e-01 -1.33210754e+00
-1.06622696e+00 -4.68084328e-02 -7.16271162e-01 5.15076339e-01
6.10669911e-01 1.96668833e-01 1.12585843e+00 -7.93639779e-01
6.34564102e-01 1.44531643e+00 4.83122349e-01 2.44639203e-01
-9.12682414e-01 -4.73231822e-01 5.21754742e-01 6.55481279e-01
-1.64208817e+00 -1.61643073e-01 9.60350394e-01 -5.62560678e-01
1.06886554e+00 1.36157706e-01 9.50252295e-01 1.10727036e+00
3.24358523e-01 9.40896630e-01 7.67032802e-01 -9.95792169e-03
-3.10273588e-01 -1.78376928e-01 -7.02549890e-02 9.20889735e-01
-4.13069785e-01 1.13458499e-01 -7.68575490e-01 1.31014571e-01
9.13720846e-01 6.86489403e-01 -5.29439807e-01 -5.32703817e-01
-1.37882888e+00 8.82745206e-01 7.82278836e-01 7.36101568e-01
-2.87693501e-01 1.44764975e-01 6.64632440e-01 4.65490580e-01
4.58815873e-01 4.58886400e-02 -4.02207702e-01 1.35335727e-02
-1.04010665e+00 6.18693465e-03 2.97498047e-01 1.01037812e+00
1.04376066e+00 -1.09923109e-01 -3.20445359e-01 7.69888759e-01
5.23830414e-01 3.13927501e-01 8.93090129e-01 -5.02924800e-01
6.68102801e-01 9.62603092e-01 -3.50330114e-01 -1.51334417e+00
-1.66758060e-01 3.19653153e-02 -7.74026394e-01 -1.77107602e-01
-6.71150237e-02 3.49552095e-01 -9.97524798e-01 1.52613437e+00
7.86471665e-02 4.09274459e-01 -1.45547003e-01 1.17755902e+00
9.21246231e-01 8.44448388e-01 -2.64798980e-02 -2.04775408e-01
1.33235800e+00 -1.14874613e+00 -8.20379019e-01 -1.60920694e-02
4.75726157e-01 -7.24828541e-01 9.97058809e-01 -1.70036554e-01
-1.01068950e+00 -7.26751089e-01 -1.03117633e+00 -4.80077207e-01
-8.86380970e-01 -6.32852688e-02 3.23038965e-01 -7.76150152e-02
-1.05076385e+00 3.16016436e-01 -8.53698373e-01 -5.43298125e-01
1.10062174e-01 2.15796083e-01 -6.38061643e-01 -4.32521105e-01
-1.48049271e+00 6.77189767e-01 4.66226101e-01 2.64688998e-01
-6.78407848e-01 -5.19931912e-01 -1.20660663e+00 -1.17433015e-02
4.50799525e-01 -6.94201291e-01 6.49615288e-01 -1.02339625e+00
-1.44197273e+00 6.90520644e-01 -2.39063889e-01 -3.77928130e-02
3.88030529e-01 -4.75330055e-02 -5.45896769e-01 6.00259185e-01
1.07952960e-01 6.06221557e-01 1.03587520e+00 -9.27176774e-01
-5.21309555e-01 -4.99070644e-01 3.65489572e-01 2.88100958e-01
-5.95837772e-01 -8.01049843e-02 -1.13398588e+00 -1.00866199e+00
1.16565123e-01 -6.11243844e-01 1.92391008e-01 2.54952490e-01
1.56636387e-02 -3.36631835e-01 1.35954607e+00 -6.71510875e-01
1.41453362e+00 -2.13739443e+00 6.33509755e-01 2.93745752e-02
3.31200927e-01 1.90304905e-01 -3.81917238e-01 4.98282313e-01
-7.01513840e-04 1.37454599e-01 1.23490028e-01 -2.61211425e-01
-6.24944679e-02 2.88936377e-01 -2.13198110e-01 4.86432612e-01
3.04772615e-01 1.23785198e+00 -1.09563196e+00 -7.18589723e-01
5.71484745e-01 7.55183995e-01 -3.88523340e-01 3.35321903e-01
-7.84499198e-02 2.37454042e-01 -8.24445724e-01 6.18403196e-01
5.18725216e-01 -2.28047475e-01 3.21516842e-02 -5.88038683e-01
-1.18170910e-01 2.39130612e-02 -8.65126193e-01 2.10537696e+00
-5.07038951e-01 7.93323159e-01 -8.13667625e-02 -1.43601108e+00
6.60080016e-01 3.60884875e-01 6.60690904e-01 -1.11930192e+00
1.35975987e-01 -3.05724367e-02 -6.17582142e-01 -1.00284696e+00
3.49344969e-01 -3.93619835e-02 3.09146028e-02 1.01353116e-01
2.76762575e-01 2.45969817e-01 -1.68831628e-02 2.54298478e-01
9.62330937e-01 4.05880213e-01 4.83198371e-03 6.55729473e-02
7.32069552e-01 -2.70568162e-01 5.36382377e-01 2.96555132e-01
-1.67955652e-01 8.96996260e-01 5.06059647e-01 -4.97931242e-01
-7.40281284e-01 -9.60369408e-01 2.20956102e-01 1.06843293e+00
5.10690451e-01 -8.20227444e-01 -1.30527005e-01 -8.24889779e-01
-7.34000057e-02 -8.53018537e-02 -9.06306505e-01 -3.49693269e-01
-5.43657780e-01 -1.57573476e-01 3.64932358e-01 6.18181884e-01
6.59628570e-01 -8.64082098e-01 -3.13719422e-01 -8.86026537e-04
-2.88828641e-01 -1.20299315e+00 -1.00563872e+00 -1.34094566e-01
-7.03055441e-01 -1.17860472e+00 -8.75368178e-01 -8.07292104e-01
3.77273798e-01 6.42935216e-01 6.66795969e-01 1.87993094e-01
-3.97456795e-01 8.15081716e-01 -6.89772844e-01 3.23371053e-01
3.91520530e-01 -1.09682038e-01 -1.24573253e-01 3.35708469e-01
4.85716254e-01 -4.60311741e-01 -7.86865771e-01 3.38761836e-01
-1.18990517e+00 -3.54449712e-02 3.96209508e-01 9.75099921e-01
5.76664388e-01 -1.62351504e-02 8.37500095e-02 -2.13658169e-01
4.99889940e-01 -5.97164214e-01 -3.35844755e-01 5.28027356e-01
-1.76726639e-01 4.00029942e-02 6.18752301e-01 -5.81326485e-01
-6.65774345e-01 -3.51885766e-01 3.08517039e-01 -1.44473851e+00
1.68827996e-01 9.03585613e-01 -1.19075812e-01 2.58480571e-02
1.09165654e-01 5.23764491e-01 -5.51218204e-02 -3.36206764e-01
3.52911174e-01 3.57696623e-01 2.01995775e-01 -4.26699460e-01
8.16348433e-01 4.62505341e-01 -9.41040665e-02 -6.68035567e-01
-6.03970289e-01 -7.58106887e-01 -8.28325868e-01 -9.61382687e-02
1.33265102e+00 -1.02220130e+00 -5.80604792e-01 3.57103050e-01
-9.54101562e-01 -2.27790669e-01 8.62239301e-02 7.60023296e-01
-4.76575077e-01 6.46932662e-01 -6.26278222e-01 -3.26437473e-01
-6.09452687e-02 -1.19184923e+00 1.29787683e+00 9.85401720e-02
2.08711460e-01 -1.17500222e+00 -1.17096893e-01 2.65562803e-01
3.77391279e-01 1.73954025e-01 9.35951829e-01 -1.94145054e-01
-8.27987313e-01 -2.32245877e-01 -5.53715110e-01 2.97876269e-01
2.56550610e-01 8.74519497e-02 -6.16295576e-01 -4.72383589e-01
-2.32044041e-01 -3.27818781e-01 1.03844297e+00 3.76737803e-01
1.25869000e+00 -2.59211868e-01 -3.58108222e-01 7.84330785e-01
1.38617074e+00 2.29753982e-02 5.06447494e-01 3.26056153e-01
1.01273584e+00 6.25077665e-01 5.99776626e-01 1.94839105e-01
5.18301904e-01 8.34319293e-01 2.50617623e-01 1.22037269e-01
-1.36231095e-01 -4.59369898e-01 4.77961868e-01 1.07002163e+00
-1.89548269e-01 -3.03377807e-01 -8.14527392e-01 6.85370684e-01
-2.27958751e+00 -1.04151595e+00 4.01863456e-01 1.97850037e+00
4.97363031e-01 -1.11798078e-01 -9.43346694e-02 -7.37238675e-02
6.27182066e-01 5.92226863e-01 -4.39689010e-01 1.39562204e-01
-9.65813994e-02 -1.93409994e-01 1.79049581e-01 4.30522472e-01
-1.08719742e+00 9.00752127e-01 5.10141754e+00 9.09796476e-01
-1.46387994e+00 1.29780114e-01 1.59175679e-01 -2.28186354e-01
-4.24574524e-01 -2.72840280e-02 -2.58405238e-01 3.74371469e-01
6.36269808e-01 3.50787165e-03 4.27534163e-01 3.44945818e-01
6.82289228e-02 4.26240772e-01 -1.08032525e+00 1.39074993e+00
3.47730547e-01 -1.31301498e+00 5.08505344e-01 2.05164067e-02
4.72041130e-01 -2.08525453e-02 7.13463277e-02 5.29360175e-01
-2.05957890e-01 -1.06913018e+00 6.93022430e-01 8.12496603e-01
9.41771328e-01 -5.59098303e-01 6.00602567e-01 1.53771862e-01
-1.95849800e+00 -1.16416320e-01 -3.84826243e-01 2.16053843e-01
4.64774854e-02 6.80263415e-02 -1.59492806e-01 9.70093369e-01
1.00763631e+00 1.39856410e+00 -5.62932968e-01 7.88067043e-01
2.97209322e-02 1.60273593e-02 -2.19735608e-01 2.05527574e-01
6.45327985e-01 -3.04320455e-01 4.74797308e-01 1.20432580e+00
2.92531461e-01 2.10363045e-01 4.26846713e-01 6.67044520e-01
-3.69879627e-03 2.23572813e-02 -7.25858867e-01 -3.70713979e-01
3.75761092e-02 1.11876416e+00 -3.99287730e-01 -2.36704633e-01
-9.55194175e-01 1.16691840e+00 2.98006952e-01 8.23978484e-01
-8.31737399e-01 -4.02008235e-01 6.91657066e-01 -8.31184816e-03
4.62774694e-01 -4.26726580e-01 3.93788964e-01 -1.64631808e+00
1.78099990e-01 -7.06021488e-01 6.17993116e-01 -9.84379172e-01
-1.47754431e+00 5.90924501e-01 3.97185460e-02 -1.48030853e+00
-1.27505064e-01 -6.69739723e-01 -4.02142584e-01 8.23613286e-01
-1.78442383e+00 -1.55142653e+00 -5.03649175e-01 1.30648112e+00
6.73311293e-01 -2.23260477e-01 5.22275209e-01 4.63231891e-01
-4.30202484e-01 6.52330041e-01 -1.57675501e-02 3.98777246e-01
7.89312899e-01 -9.27360415e-01 -2.49007896e-01 5.69012403e-01
1.34675294e-01 7.23439097e-01 8.07931498e-02 -3.34248900e-01
-1.76153314e+00 -1.16609192e+00 6.39106989e-01 -2.09182546e-01
8.86166394e-01 -3.11508805e-01 -1.01493585e+00 6.96520984e-01
2.40099743e-01 5.86852551e-01 4.74706411e-01 -4.81193177e-02
-6.94613636e-01 -2.50255376e-01 -7.34525025e-01 6.02831125e-01
1.07748425e+00 -1.31616867e+00 -7.17238188e-01 7.70114735e-02
8.95015955e-01 -1.92780212e-01 -1.00708616e+00 4.51490790e-01
6.58048391e-01 -8.18586111e-01 1.05521393e+00 -6.19550943e-01
5.11455357e-01 -4.64421570e-01 -4.16386336e-01 -9.17830050e-01
-3.52127820e-01 -3.28671664e-01 -1.65816009e-01 1.13415730e+00
1.95900258e-02 -4.37189758e-01 3.38074386e-01 9.40159783e-02
1.20684765e-02 -9.34074104e-01 -9.76674199e-01 -6.34149253e-01
-1.09701127e-01 -4.08670843e-01 4.11909461e-01 1.21294951e+00
1.07539855e-02 3.38804960e-01 -4.26867992e-01 1.55817002e-01
1.01465270e-01 2.25131616e-01 5.06498396e-01 -9.41440642e-01
-1.03782713e-01 -5.87764442e-01 -8.66022944e-01 -1.44514334e+00
3.24736595e-01 -9.10200655e-01 -1.57745704e-01 -1.50585997e+00
3.17272216e-01 -3.51138748e-02 -7.52880514e-01 4.41104472e-01
-7.14436099e-02 3.04096997e-01 2.06311047e-01 4.20710027e-01
-7.58697867e-01 1.00383818e+00 1.39919412e+00 -5.28249145e-01
-1.23578653e-01 -5.73681891e-01 -1.87263280e-01 4.65402901e-01
2.43499562e-01 -6.79468140e-02 -6.63792729e-01 -7.37634659e-01
2.53508806e-01 3.00536871e-01 5.97648442e-01 -8.43156755e-01
4.70515668e-01 -3.64416659e-01 5.32636821e-01 -6.07194424e-01
5.53934813e-01 -1.00391257e+00 -1.61768302e-01 5.45241944e-02
-3.33372384e-01 2.95280665e-01 8.01534504e-02 1.00761461e+00
-9.15579140e-01 2.02787846e-01 4.10732776e-01 -1.02196284e-01
-9.58162189e-01 1.01114559e+00 -2.03066185e-01 -1.91199616e-01
1.07797217e+00 -4.73860145e-01 -2.03293607e-01 -3.55931848e-01
-6.75236583e-01 4.63005900e-01 4.46798801e-01 1.00480843e+00
8.79015207e-01 -1.67391348e+00 -2.68079817e-01 3.04838628e-01
4.88935828e-01 -2.13468477e-01 6.26560032e-01 8.94924939e-01
-4.33898121e-01 6.08825028e-01 -1.77939162e-01 -8.56396556e-01
-1.13073349e+00 9.21874166e-01 4.60647643e-01 -2.42791995e-01
-7.02767193e-01 7.27440596e-01 4.88180161e-01 -2.22574696e-01
9.93833989e-02 -2.51057476e-01 -4.07775700e-01 3.85526836e-01
3.84648353e-01 -6.15995415e-02 -2.19849288e-01 -9.28050816e-01
-4.38424706e-01 1.23303878e+00 -4.33850847e-02 3.64258550e-02
1.14056420e+00 -4.36864227e-01 -2.23866507e-01 5.89445889e-01
1.98496604e+00 -3.06237429e-01 -1.19541419e+00 -6.93945587e-01
-2.75909215e-01 -7.68752396e-01 3.78195077e-01 -1.79553181e-01
-1.40442419e+00 1.27797043e+00 5.42411566e-01 2.37992674e-01
1.23225760e+00 2.31687389e-02 8.91423464e-01 1.89873740e-01
6.90090377e-03 -1.03514147e+00 5.95360935e-01 5.64763427e-01
1.00764322e+00 -1.23992527e+00 -7.43769333e-02 -1.05146579e-01
-4.32020783e-01 1.36103165e+00 6.86150789e-01 -1.28314868e-01
9.53933299e-01 -3.38383913e-01 -1.35525092e-01 -3.99931908e-01
-8.46873641e-01 -3.16617429e-01 7.98658431e-01 3.59930217e-01
3.83751869e-01 -2.86956996e-01 -1.03813984e-01 3.42012435e-01
3.57479572e-01 -1.01107985e-01 -1.97262019e-01 9.36488271e-01
-9.22158062e-02 -9.70405817e-01 -1.75569057e-02 2.42600247e-01
-2.12767109e-01 -1.24394624e-02 -2.00939685e-01 8.45465779e-01
1.97698399e-01 7.37261534e-01 2.59917259e-01 -6.37602985e-01
2.74570912e-01 -3.86298075e-02 4.01543856e-01 -3.95201683e-01
-2.15194419e-01 4.16881084e-01 -4.79138404e-01 -8.29432249e-01
-7.57998884e-01 -3.86933804e-01 -1.03597379e+00 -2.90299773e-01
-1.47092879e-01 1.53370202e-02 2.26053670e-01 9.97353733e-01
4.78328884e-01 6.21113181e-01 5.45300245e-01 -1.03362656e+00
-2.20059287e-02 -8.52164745e-01 -5.52661002e-01 5.18364668e-01
6.91780686e-01 -7.82438636e-01 -3.14460754e-01 4.98695299e-02]
|
[10.337059020996094, 0.9699962735176086]
|
bde7c416-b369-486b-8141-c058a2a2ec39
|
generating-3d-adversarial-point-clouds
|
1809.07016
| null |
https://arxiv.org/abs/1809.07016v4
|
https://arxiv.org/pdf/1809.07016v4.pdf
|
Generating 3D Adversarial Point Clouds
|
Deep neural networks are known to be vulnerable to adversarial examples which are carefully crafted instances to cause the models to make wrong predictions. While adversarial examples for 2D images and CNNs have been extensively studied, less attention has been paid to 3D data such as point clouds. Given many safety-critical 3D applications such as autonomous driving, it is important to study how adversarial point clouds could affect current deep 3D models. In this work, we propose several novel algorithms to craft adversarial point clouds against PointNet, a widely used deep neural network for point cloud processing. Our algorithms work in two ways: adversarial point perturbation and adversarial point generation. For point perturbation, we shift existing points negligibly. For point generation, we generate either a set of independent and scattered points or a small number (1-3) of point clusters with meaningful shapes such as balls and airplanes which could be hidden in the human psyche. In addition, we formulate six perturbation measurement metrics tailored to the attacks in point clouds and conduct extensive experiments to evaluate the proposed algorithms on the ModelNet40 3D shape classification dataset. Overall, our attack algorithms achieve a success rate higher than 99% for all targeted attacks
|
['Chong Xiang', 'Bo Li', 'Charles R. Qi']
|
2018-09-19
|
generating-3d-adversarial-point-clouds-1
|
http://openaccess.thecvf.com/content_CVPR_2019/html/Xiang_Generating_3D_Adversarial_Point_Clouds_CVPR_2019_paper.html
|
http://openaccess.thecvf.com/content_CVPR_2019/papers/Xiang_Generating_3D_Adversarial_Point_Clouds_CVPR_2019_paper.pdf
|
cvpr-2019-6
|
['3d-shape-retrieval']
|
['computer-vision']
|
[ 2.58664563e-02 1.97026789e-01 4.26927388e-01 -2.14947239e-01
-4.72330213e-01 -9.83463049e-01 7.19882607e-01 2.23516688e-01
-1.81982875e-01 3.35930169e-01 -4.90646452e-01 -5.70127547e-01
2.35624090e-01 -1.06657541e+00 -1.38420665e+00 -7.79705286e-01
-1.59624875e-01 5.08442044e-01 3.58542681e-01 -4.83755171e-01
3.46855104e-01 1.36376882e+00 -1.23429918e+00 -5.83431907e-02
7.42648005e-01 9.47085857e-01 -5.09336770e-01 5.54980993e-01
-1.97769832e-02 4.58382994e-01 -1.15333009e+00 -7.50939548e-01
8.74416769e-01 2.67162085e-01 -2.79824615e-01 -2.82107681e-01
5.75310647e-01 -3.36721301e-01 -5.74596524e-01 1.54828036e+00
5.29740989e-01 1.37999102e-01 6.89212918e-01 -1.71414864e+00
-8.50521803e-01 3.35192978e-01 -6.11131608e-01 5.39048277e-02
5.24689490e-03 6.23861134e-01 1.57675147e-01 -6.29928231e-01
6.69924766e-02 1.57873118e+00 7.06784427e-01 7.38917589e-01
-8.82970512e-01 -1.27442789e+00 2.98144743e-02 8.48440826e-02
-1.27445650e+00 -5.84221184e-02 1.07967913e+00 -3.99716735e-01
7.26493061e-01 5.24299145e-01 4.48970914e-01 1.54626811e+00
5.54858506e-01 4.40641522e-01 7.14233518e-01 1.44679412e-01
3.15601528e-01 4.32175547e-02 -1.81345448e-01 6.51782900e-02
5.09204090e-01 6.00636542e-01 1.14942744e-01 -5.55272877e-01
6.82012975e-01 2.70985752e-01 -9.44977552e-02 -3.00336719e-01
-1.10076511e+00 9.02613342e-01 9.81539130e-01 -2.05761626e-01
-4.54655617e-01 4.32198197e-01 4.87429202e-01 2.16739163e-01
3.44797581e-01 5.32790124e-01 -2.32168749e-01 3.04202616e-01
-1.52466804e-01 8.02452981e-01 4.74862009e-01 1.30274868e+00
4.88231510e-01 3.49081606e-01 -3.75469439e-02 3.73860687e-01
1.63367853e-01 1.09242892e+00 2.18968075e-02 -6.46122575e-01
6.17088675e-01 4.90960240e-01 2.34848246e-01 -1.66884708e+00
-1.27129644e-01 -2.20541224e-01 -1.21507823e+00 7.85821021e-01
2.09580600e-01 -7.45081678e-02 -1.07748115e+00 1.52394128e+00
5.38566709e-01 5.60257316e-01 1.46794200e-01 1.02602220e+00
6.88857019e-01 6.23211920e-01 6.44852519e-02 4.74098712e-01
9.83464360e-01 -3.83978993e-01 -2.33670205e-01 -2.83273011e-02
2.35570073e-01 -6.93144679e-01 8.05960357e-01 2.15562776e-01
-1.08929396e+00 -6.25034273e-01 -1.22527218e+00 2.75380731e-01
-5.64598680e-01 -8.00098479e-01 2.24130169e-01 8.77227902e-01
-4.66224313e-01 7.91888416e-01 -9.40704703e-01 1.51397899e-01
7.97677696e-01 4.31364924e-01 -4.61336255e-01 2.93615442e-02
-1.42204976e+00 1.09431732e+00 2.53562659e-01 7.15451986e-02
-1.14203894e+00 -9.46188748e-01 -7.01931596e-01 -1.45363808e-01
-5.41331545e-02 -6.70843780e-01 9.21545565e-01 -4.98766631e-01
-1.05021644e+00 8.82576704e-01 2.97302425e-01 -9.41650629e-01
6.69780493e-01 -2.39131674e-01 -4.83736187e-01 1.26351714e-01
-1.96558341e-01 6.48038089e-01 1.22999597e+00 -1.58627975e+00
-3.40508372e-01 -6.59633160e-01 3.48563969e-01 2.01197546e-02
-1.37923092e-01 -8.95685852e-02 3.14267948e-02 -8.10977697e-01
9.56466198e-02 -1.33130002e+00 -4.99758512e-01 2.03224376e-01
-7.86981702e-01 7.72212865e-03 1.12793934e+00 5.09773754e-02
2.23138690e-01 -2.26800227e+00 -1.44817024e-01 2.89353639e-01
3.85737568e-01 5.19028842e-01 -9.49760526e-02 1.51343778e-01
-4.61601555e-01 4.90586162e-01 -2.13509619e-01 -1.50194719e-01
2.59062886e-01 3.50023568e-01 -9.75050747e-01 8.60625625e-01
5.20740032e-01 9.09775436e-01 -8.32994640e-01 1.19468026e-01
4.12946761e-01 4.21064079e-01 -6.10576570e-01 1.36095941e-01
-1.53939024e-01 4.79119450e-01 -7.01228619e-01 6.73422933e-01
1.32868814e+00 4.18861747e-01 -1.05824792e+00 7.78449047e-03
4.80771959e-01 -6.61820322e-02 -8.67399156e-01 1.05341637e+00
-1.10331066e-01 3.10075581e-01 -2.66237259e-01 -7.13661253e-01
1.22710931e+00 1.89976230e-01 2.97269762e-01 -1.49188533e-01
3.47228289e-01 1.02414206e-01 8.92061293e-02 -1.20206133e-01
4.31066304e-01 -3.50372821e-01 -3.54287893e-01 9.02292132e-02
-4.97129947e-01 -8.00163031e-01 -7.98941135e-01 -2.58480338e-03
1.32793581e+00 -3.61571908e-01 -1.42449081e-01 9.86085758e-02
4.99716222e-01 2.47835532e-01 4.90511298e-01 7.98435032e-01
-4.70310420e-01 7.32123256e-01 3.73492032e-01 -6.45774841e-01
-1.32966888e+00 -1.41799593e+00 -1.17836736e-01 1.86119929e-01
5.98784387e-01 1.83703884e-01 -5.73918581e-01 -8.14718366e-01
4.84801799e-01 1.16888261e+00 -7.23869324e-01 -1.02476895e+00
-5.58522642e-01 -3.30660790e-01 1.12989807e+00 4.28010195e-01
5.28308630e-01 -1.11371374e+00 -2.75366157e-01 -1.18802652e-01
4.17201757e-01 -1.14959586e+00 -2.26032078e-01 6.83781281e-02
-7.10287273e-01 -9.87472177e-01 -4.56866413e-01 -4.59507883e-01
7.40767062e-01 3.10627699e-01 1.02819049e+00 -5.58286346e-02
1.37666851e-01 1.51205197e-01 -5.00015676e-01 -1.28554285e+00
-9.24307942e-01 -2.86825627e-01 5.94191253e-01 -1.34577975e-01
6.98933244e-01 -7.66321480e-01 -5.23216307e-01 5.27965605e-01
-1.07848406e+00 -5.97254574e-01 3.74266744e-01 4.81485128e-01
5.85209608e-01 2.74194151e-01 1.98718622e-01 -6.40062153e-01
6.00929141e-01 -6.46296144e-01 -5.91387987e-01 -3.86871040e-01
8.08774903e-02 -2.71734625e-01 9.67339575e-01 -8.46835375e-01
-1.56072304e-01 -4.56815325e-02 -4.72347349e-01 -1.51761532e+00
-5.59652388e-01 -1.17046647e-01 -3.89625043e-01 -6.80429518e-01
1.03959858e+00 1.19852386e-01 3.88795212e-02 -9.48399678e-02
4.06077713e-01 3.04776967e-01 6.31192088e-01 -6.33907437e-01
1.78235281e+00 6.20484471e-01 2.60974586e-01 -6.65507615e-01
-3.93855482e-01 6.61386922e-02 -3.73761564e-01 -8.69733915e-02
7.42019057e-01 -7.14613378e-01 -9.31449175e-01 6.04654729e-01
-1.48790622e+00 1.11372270e-01 -2.44730696e-01 1.72594592e-01
-5.81292927e-01 3.72285515e-01 -1.40351728e-01 -6.06074095e-01
-2.84528077e-01 -1.46283841e+00 1.07950091e+00 -1.23702981e-01
3.43163237e-02 -5.81542492e-01 -1.86486781e-01 -9.31683853e-02
1.94692403e-01 1.01781034e+00 8.67697537e-01 -1.09884608e+00
-6.51744604e-01 -7.98219442e-01 5.05375229e-02 5.64883351e-01
9.69346613e-03 3.64182033e-02 -9.10625935e-01 -3.20900291e-01
4.30274904e-01 -2.03280121e-01 3.19983393e-01 -4.13711108e-02
1.50025165e+00 -4.88334149e-01 -2.65736848e-01 8.88814151e-01
9.62148309e-01 4.03713167e-01 7.45429873e-01 3.54956716e-01
9.06661153e-01 2.63404638e-01 4.81048703e-01 2.03925520e-01
-2.82069236e-01 4.79453355e-01 1.39304233e+00 1.29636630e-01
5.06905854e-01 -3.46268326e-01 8.03011432e-02 2.11443603e-01
2.85441447e-02 -2.85858899e-01 -9.85781252e-01 3.80455166e-01
-1.41043437e+00 -1.00422955e+00 -1.99788392e-01 2.04808950e+00
4.48503315e-01 7.36327410e-01 -1.27817512e-01 4.36054587e-01
8.65839541e-01 1.95234075e-01 -9.46574271e-01 -4.63763624e-01
-2.42660996e-02 2.62035310e-01 9.57352519e-01 1.53221888e-02
-1.23683333e+00 9.61845517e-01 4.92379665e+00 7.51620650e-01
-1.12850678e+00 -1.79808944e-01 4.85965610e-01 -2.65361786e-01
-2.88939506e-01 -2.87617147e-01 -6.09224021e-01 5.81852436e-01
6.76836729e-01 -2.99971789e-01 1.41629621e-01 1.22347128e+00
7.55376667e-02 9.15778458e-01 -1.05172455e+00 1.01833558e+00
-2.64600873e-01 -1.32401919e+00 5.87051809e-01 2.85825551e-01
6.26325071e-01 2.01535016e-01 5.40260375e-01 3.17617357e-01
5.97463310e-01 -1.22948933e+00 8.41147304e-01 2.44231403e-01
4.46898401e-01 -1.28848386e+00 5.82608283e-01 6.61609888e-01
-6.10522628e-01 1.02963880e-01 -9.58525896e-01 2.34384373e-01
-7.63429375e-03 4.03321087e-01 -8.61375093e-01 4.01161790e-01
8.80606592e-01 4.32753474e-01 -2.47301355e-01 9.24185455e-01
-2.17336625e-01 3.84050667e-01 -5.84857285e-01 -1.62874646e-02
6.74804211e-01 1.18076585e-01 1.25753188e+00 5.06023467e-01
2.76688874e-01 2.69937813e-01 -1.15310729e-01 1.18627560e+00
-2.20229849e-01 -3.62399608e-01 -1.33728075e+00 1.76897824e-01
7.44847476e-01 8.60940337e-01 -2.93302476e-01 4.45647687e-02
6.92770779e-02 8.13125372e-01 3.35745811e-02 3.27543825e-01
-1.18820393e+00 -5.44733226e-01 1.44936931e+00 1.19526625e-01
2.54259080e-01 -4.30122137e-01 -5.24899065e-01 -7.38929152e-01
3.25860754e-02 -9.39824760e-01 -2.48061031e-01 -7.36805797e-01
-1.68070650e+00 8.43033969e-01 -9.27754864e-02 -1.68643785e+00
5.17543517e-02 -7.02010572e-01 -9.54068244e-01 1.00583386e+00
-1.13510048e+00 -9.81498361e-01 -3.14017147e-01 8.71637523e-01
1.27083018e-01 -4.80885714e-01 6.66319728e-01 6.15995638e-02
-2.67808229e-01 8.53649497e-01 -1.06955789e-01 3.60276878e-01
3.73096347e-01 -1.05020678e+00 1.51510668e+00 8.75555098e-01
-1.38324639e-02 6.31196916e-01 9.39241469e-01 -6.92486167e-01
-1.38990164e+00 -1.65022278e+00 1.14329509e-01 -9.09119487e-01
6.10174894e-01 -4.01602775e-01 -1.22512829e+00 6.60835505e-01
-3.37676764e-01 3.00622523e-01 2.78888524e-01 -5.35917640e-01
-4.28088069e-01 5.24026854e-03 -1.67296624e+00 8.81767631e-01
9.24058437e-01 -2.64095217e-01 -7.28082001e-01 5.99181890e-01
1.30257761e+00 -9.40036476e-01 -8.43115747e-01 7.31882691e-01
-1.05785780e-01 -8.94642413e-01 1.52693534e+00 -9.71513987e-01
4.94468868e-01 -4.51057196e-01 -2.58150190e-01 -1.47053671e+00
-2.31253907e-01 -6.15859449e-01 -5.30441590e-02 7.18140662e-01
8.87183025e-02 -8.67137074e-01 9.40069497e-01 5.04597664e-01
-4.79310125e-01 -6.59040868e-01 -1.17917633e+00 -9.89567161e-01
6.54772878e-01 -7.23669946e-01 1.26808250e+00 1.00384891e+00
-7.54478812e-01 -2.47416452e-01 -1.96441695e-01 1.11056423e+00
8.10040712e-01 -2.59449005e-01 1.28509033e+00 -1.00989878e+00
1.65149137e-01 -5.83512187e-01 -1.10845602e+00 -7.73528337e-01
2.96331167e-01 -7.14540780e-01 -1.14878416e-01 -7.65630901e-01
-6.47566915e-01 -8.19164276e-01 -1.72705248e-01 2.62947321e-01
-1.86613634e-01 4.05398309e-01 3.40608776e-01 1.98176384e-01
1.02911621e-01 7.13046193e-01 1.29459667e+00 -5.80971241e-01
1.48939684e-01 6.01495504e-01 -6.26416862e-01 9.09909785e-01
7.79154122e-01 -8.08596909e-01 -3.72347653e-01 -4.99280810e-01
1.26214221e-01 -1.37213260e-01 8.63149583e-01 -1.27413797e+00
-1.09299342e-03 -2.38564983e-01 4.47368532e-01 -9.38784838e-01
5.37352026e-01 -1.32111943e+00 1.52749896e-01 4.80096847e-01
4.60435674e-02 1.73327863e-01 6.21139884e-01 8.08790386e-01
-7.56330192e-02 -7.45952427e-02 1.09738159e+00 -1.22774340e-01
-2.88632661e-01 8.40167999e-01 1.04421526e-01 -1.38553441e-01
1.51291752e+00 -3.21205854e-01 -3.25980514e-01 -2.49852568e-01
-5.54952323e-01 1.50483474e-01 7.13364899e-01 8.41913342e-01
9.16872084e-01 -1.70114994e+00 -7.54916251e-01 3.77586722e-01
1.80463374e-01 6.77385509e-01 1.74284726e-01 9.63092223e-02
-8.36289108e-01 1.36405006e-01 -3.76274347e-01 -8.63608599e-01
-9.42010343e-01 1.15342045e+00 5.41667581e-01 4.36676919e-01
-7.85339952e-01 8.97954464e-01 3.04974884e-01 -6.02611780e-01
2.66438931e-01 -4.92086411e-01 1.47633180e-01 -6.83534086e-01
3.43682259e-01 5.66097312e-02 1.60863549e-01 -6.88594222e-01
-3.99782300e-01 3.07765037e-01 -5.25137149e-02 5.64440191e-01
1.10703397e+00 4.56434011e-01 1.35219604e-01 1.34209946e-01
1.24748480e+00 -7.61210993e-02 -1.15187442e+00 1.11042425e-01
-4.84855503e-01 -5.38799107e-01 -2.87461728e-01 -2.90024966e-01
-9.89467978e-01 9.64668512e-01 4.90353137e-01 4.98948246e-01
7.87713408e-01 -2.68849730e-01 1.08664143e+00 5.63207209e-01
4.78754848e-01 -1.33662492e-01 -3.68718430e-02 5.80121577e-01
1.24768507e+00 -1.14331770e+00 -2.35722631e-01 -2.84446537e-01
-6.19600117e-01 7.01370239e-01 6.53477967e-01 -8.89353335e-01
7.57616639e-01 1.49041235e-01 1.24580435e-01 -1.52025238e-01
-4.64883059e-01 4.73761678e-01 2.90983498e-01 9.07330692e-01
-5.61295629e-01 -5.54340258e-02 4.93757397e-01 6.03653073e-01
-7.53239155e-01 -5.04475057e-01 4.06042546e-01 7.59983659e-01
-3.44485313e-01 -7.02440202e-01 -9.36110556e-01 2.13104099e-01
-3.00009161e-01 1.10263526e-01 -4.63711441e-01 9.69275415e-01
1.50274903e-01 6.27582073e-01 1.96403354e-01 -7.82510281e-01
8.07588279e-01 -2.65330642e-01 1.90840915e-01 -4.28235769e-01
-7.01926351e-01 -6.26578808e-01 -4.37372863e-01 -5.31991601e-01
1.23956472e-01 -4.13095206e-01 -1.10134172e+00 -7.71072984e-01
-2.57911403e-02 -8.60267803e-02 7.48452902e-01 6.65903747e-01
4.29875106e-01 4.49507803e-01 9.43751931e-01 -1.27763641e+00
-1.08123767e+00 -8.23226690e-01 -4.52136874e-01 7.66338050e-01
5.04918396e-01 -9.45307672e-01 -6.85951948e-01 -4.08901513e-01]
|
[7.7050323486328125, -4.4678239822387695]
|
c6758ee9-abc3-4e82-bc40-88cfb8eafff1
|
exploring-hierarchy-aware-inverse
|
1807.05037
| null |
http://arxiv.org/abs/1807.05037v1
|
http://arxiv.org/pdf/1807.05037v1.pdf
|
Exploring Hierarchy-Aware Inverse Reinforcement Learning
|
We introduce a new generative model for human planning under the Bayesian
Inverse Reinforcement Learning (BIRL) framework which takes into account the
fact that humans often plan using hierarchical strategies. We describe the
Bayesian Inverse Hierarchical RL (BIHRL) algorithm for inferring the values of
hierarchical planners, and use an illustrative toy model to show that BIHRL
retains accuracy where standard BIRL fails. Furthermore, BIHRL is able to
accurately predict the goals of `Wikispeedia' game players, with inclusion of
hierarchical structure in the model resulting in a large boost in accuracy. We
show that BIHRL is able to significantly outperform BIRL even when we only have
a weak prior on the hierarchical structure of the plans available to the agent,
and discuss the significant challenges that remain for scaling up this
framework to more realistic settings.
|
['Daniel Filan', 'Chris Cundy']
|
2018-07-13
| null | null | null | null |
['birl-cima']
|
['medical']
|
[ 1.02103293e-01 9.15366292e-01 1.00411676e-01 -1.47742769e-02
-8.37148011e-01 -5.76125383e-01 8.66811156e-01 -1.27917469e-01
-4.93463159e-01 1.08268511e+00 7.70627677e-01 -1.84148803e-01
-6.07671320e-01 -8.07572901e-01 -7.07669020e-01 -5.30449808e-01
-2.06995040e-01 1.15381789e+00 3.76591593e-01 -4.20531660e-01
2.30424747e-01 1.71751872e-01 -1.21901441e+00 1.17575884e-01
2.70979702e-01 2.27076292e-01 3.25446039e-01 1.14281321e+00
7.69964993e-01 1.44094074e+00 -2.56350219e-01 -2.48687062e-02
3.70288044e-01 -5.29877543e-01 -1.24853516e+00 1.87228818e-03
-4.95562464e-01 -8.31041992e-01 -4.99522746e-01 6.78368509e-01
3.91946346e-01 3.41427654e-01 7.90615320e-01 -1.22444296e+00
-5.88579737e-02 1.00672293e+00 -2.14981928e-01 3.02477628e-02
7.86937833e-01 2.63124466e-01 1.32981050e+00 -4.63470593e-02
6.02346599e-01 1.65785050e+00 5.95903814e-01 3.69955361e-01
-1.29685509e+00 -3.35941672e-01 3.02865028e-01 3.06803316e-01
-1.09690475e+00 -2.97706753e-01 3.23606431e-01 -3.40160459e-01
1.05760670e+00 -8.63029659e-02 8.45907271e-01 8.92496943e-01
1.78274736e-01 8.32082748e-01 1.47118437e+00 -4.67763782e-01
3.91823560e-01 -4.20287341e-01 1.43618211e-01 8.92777383e-01
3.76670342e-03 7.01082647e-01 -7.43425369e-01 -2.55776793e-01
1.15834439e+00 -4.58888680e-01 1.86410733e-02 -4.69170898e-01
-1.14436316e+00 9.78533268e-01 3.07731122e-01 -7.48674348e-02
-6.02660716e-01 5.53382516e-01 3.96283856e-03 1.33625463e-01
-1.88332796e-01 4.62592810e-01 -1.74939692e-01 -3.75500500e-01
-3.74871850e-01 7.31213927e-01 9.03360486e-01 1.05019021e+00
6.78263128e-01 -4.00144130e-01 -1.60447136e-01 2.17040330e-01
5.46320975e-01 8.85517299e-02 1.31266415e-01 -1.87244678e+00
2.45159313e-01 1.32428452e-01 5.68891943e-01 -6.64014518e-01
-6.77254140e-01 -1.78248435e-01 -1.83018401e-01 2.58178979e-01
5.35272479e-01 -1.65593967e-01 -8.20605338e-01 1.94600487e+00
2.35752881e-01 -7.05372244e-02 3.26011837e-01 4.64660406e-01
3.45246017e-01 6.26381636e-01 1.99011015e-03 -2.40965828e-01
1.04027438e+00 -7.07266331e-01 -4.79460239e-01 -5.46370149e-01
6.71105266e-01 -5.46418689e-02 8.76631200e-01 6.12905562e-01
-1.26924872e+00 -8.22044015e-02 -1.03499496e+00 -1.43472970e-01
2.41822913e-01 -4.52971488e-01 1.08724642e+00 4.89624381e-01
-1.29078054e+00 6.49609208e-01 -1.20549417e+00 -3.03023636e-01
2.30686650e-01 6.30042732e-01 -1.07671186e-01 -2.58364856e-01
-1.17136478e+00 1.13938367e+00 7.15665162e-01 -1.78293705e-01
-1.43002260e+00 -1.06358744e-01 -9.54893708e-01 1.69152707e-01
9.03496325e-01 -6.90588593e-01 1.89256048e+00 -2.89547950e-01
-1.81283069e+00 2.47912079e-01 5.79224601e-02 -5.96493483e-01
5.53841352e-01 -3.48935612e-02 4.82442796e-01 1.06863067e-01
3.49560648e-01 7.92286634e-01 1.91926643e-01 -1.21705031e+00
-8.10920715e-01 -4.09461558e-01 6.84322417e-01 7.70018935e-01
5.09277165e-01 -2.32149675e-01 -1.67636499e-01 2.68676132e-02
1.95184022e-01 -1.17039108e+00 -6.34203136e-01 -7.50757158e-01
-3.43047172e-01 -4.87225413e-01 -6.70560971e-02 -4.85445142e-01
8.73706698e-01 -1.67490208e+00 4.70863193e-01 2.44275019e-01
2.52080262e-01 -4.96592969e-01 -9.98752117e-02 6.85333073e-01
2.56060749e-01 -2.03503713e-01 -1.10146418e-01 -1.49746373e-01
4.23005074e-01 6.02255702e-01 -1.58558369e-01 3.35380793e-01
-5.08907557e-01 1.02756834e+00 -1.11895823e+00 -4.83552039e-01
1.88773245e-01 7.37100765e-02 -1.03429902e+00 3.47558558e-01
-3.11925083e-01 5.68599820e-01 -6.68457508e-01 5.36816940e-02
-1.30555347e-01 -1.20546848e-01 6.76440954e-01 5.06011069e-01
-2.29249354e-02 6.02915049e-01 -1.18015540e+00 1.31078565e+00
-1.74268320e-01 2.79092267e-02 9.24062580e-02 -6.05323672e-01
1.96003929e-01 3.62813592e-01 3.81385624e-01 -4.19902861e-01
5.65060861e-02 -1.68025121e-01 3.50133151e-01 -9.78582650e-02
1.63190484e-01 -5.60692489e-01 -4.17579502e-01 7.68195212e-01
-8.68439209e-03 -5.34735799e-01 2.51546115e-01 3.91324610e-01
1.26222920e+00 5.38822651e-01 7.91097343e-01 -1.63922161e-01
1.22648582e-01 1.72902137e-01 6.45920992e-01 1.37896121e+00
-1.92394644e-01 3.19218338e-02 9.23327148e-01 -3.50063682e-01
-7.99085796e-01 -9.58358645e-01 3.75105977e-01 1.48153245e+00
1.42399520e-01 -4.41603452e-01 -8.67020369e-01 -4.67107743e-01
-4.07959938e-01 1.12141550e+00 -8.04385364e-01 1.91813111e-02
-5.20115316e-01 -9.61668193e-01 5.15143633e-01 6.41208529e-01
4.18677628e-01 -1.30797720e+00 -9.94707406e-01 3.23375970e-01
-2.28411570e-01 -1.02364421e+00 -2.19487086e-01 3.69684190e-01
-6.79151773e-01 -1.06984103e+00 -6.78129271e-02 -3.74066859e-01
1.96533158e-01 -1.39152408e-01 1.16278458e+00 -3.44884098e-02
1.70367509e-01 7.93899536e-01 -3.45641434e-01 -2.96880871e-01
-6.37630522e-01 -1.39123304e-02 -9.16995257e-02 -7.49992251e-01
-1.42232686e-01 -6.96446240e-01 -3.38879079e-01 1.06735118e-01
-5.11896908e-01 6.03984118e-01 4.18212980e-01 8.97330880e-01
1.00271367e-01 3.24885666e-01 2.73481816e-01 -1.11987412e+00
8.31368506e-01 -2.49474674e-01 -8.40902746e-01 -9.55911502e-02
-6.40723586e-01 4.19775397e-01 1.12613447e-01 1.83903221e-02
-1.18186498e+00 1.27559483e-01 -1.86609179e-01 3.02592069e-01
-2.41392523e-01 4.62543815e-01 -1.36539206e-01 1.42622635e-01
5.47457576e-01 -2.15090364e-02 -1.58789121e-02 -1.65977925e-01
4.97422218e-01 2.63928860e-01 4.58312571e-01 -8.77113044e-01
4.30585563e-01 2.81531245e-01 2.85010010e-01 -4.82110232e-01
-1.07436347e+00 -1.37417549e-02 -6.21647775e-01 -2.07201391e-01
9.29666519e-01 -8.62470686e-01 -1.47409356e+00 4.68960032e-02
-9.38770890e-01 -1.10271680e+00 -2.92262793e-01 4.61383134e-01
-1.65650761e+00 2.06705570e-01 -9.55574691e-01 -1.07447720e+00
4.17260766e-01 -1.47262478e+00 8.95290136e-01 -1.03197560e-01
-3.92672896e-01 -9.41211462e-01 2.91885167e-01 4.20408279e-01
-9.56056118e-02 2.85890624e-02 1.12401807e+00 -6.39084756e-01
-9.35257316e-01 2.72717863e-01 1.43164083e-01 -4.38445479e-01
-1.77187994e-01 -6.17903769e-01 -6.73553765e-01 -2.15535685e-01
6.79554567e-02 -5.93123376e-01 5.54887891e-01 6.32480919e-01
4.86522436e-01 -5.77985764e-01 -3.31489742e-01 1.48525029e-01
1.09593213e+00 4.00294334e-01 7.52429664e-01 5.28385282e-01
4.62459892e-01 6.66998267e-01 6.85318470e-01 5.93933046e-01
9.90109265e-01 7.62109578e-01 5.73571920e-01 3.82102013e-01
2.29079366e-01 -6.11209035e-01 3.87068868e-01 3.71853173e-01
-4.45382535e-01 -5.21483198e-02 -8.54749918e-01 3.93832892e-01
-2.19300413e+00 -1.09184003e+00 1.76746845e-01 2.03696823e+00
9.11724389e-01 2.76385844e-01 5.41594982e-01 6.46254420e-02
4.23231751e-01 -1.45178847e-02 -4.97534424e-01 -3.14054906e-01
5.17723799e-01 2.12105572e-01 6.55495703e-01 1.28625786e+00
-8.72847140e-01 1.28922129e+00 8.13837814e+00 4.14473802e-01
1.90369725e-01 1.83312297e-01 2.65446156e-01 2.39441425e-01
1.06149074e-02 3.78871590e-01 -7.25107133e-01 -7.24984258e-02
1.03686154e+00 -1.98582664e-01 1.14018309e+00 5.70008218e-01
6.93221018e-02 -5.49792945e-01 -1.37855947e+00 4.72085506e-01
-2.66055703e-01 -9.91947293e-01 -2.97315955e-01 4.66378927e-01
4.55626398e-01 -2.29942836e-02 -2.79792041e-01 7.72276998e-01
1.61508691e+00 -1.22836530e+00 7.73951709e-01 3.27646494e-01
2.41755217e-01 -9.78870809e-01 5.28744459e-01 1.12306273e+00
-7.55067110e-01 -4.89064515e-01 -2.27797493e-01 -6.88334525e-01
3.21128935e-01 -3.33464622e-01 -1.40280426e+00 2.79471785e-01
4.80445355e-01 5.52773237e-01 -2.37887457e-01 5.04086792e-01
-6.15718246e-01 3.98208439e-01 -3.41351479e-01 6.88915402e-02
4.90853459e-01 -2.18994901e-01 3.92120808e-01 7.39441693e-01
-4.09311727e-02 6.88422680e-01 6.58613503e-01 7.61140645e-01
4.49307144e-01 -5.11441171e-01 -5.91199934e-01 -1.15163185e-01
2.85449535e-01 7.55597949e-01 -7.72776306e-01 -2.35230803e-01
1.41550884e-01 6.35011315e-01 6.04098856e-01 1.17883042e-01
-6.50209069e-01 3.52639377e-01 3.62715125e-01 -5.47918566e-02
9.04207826e-02 -4.52848643e-01 -1.21887714e-01 -6.60382271e-01
-7.50697136e-01 -1.01328743e+00 5.82850099e-01 -9.48852360e-01
-7.76755154e-01 3.18942189e-01 5.24316788e-01 -5.12949407e-01
-1.09763753e+00 -5.57036221e-01 3.85863297e-02 5.74083149e-01
-1.00006783e+00 -1.04450345e+00 1.06370477e-02 3.92345935e-01
5.78357577e-01 2.80743927e-01 9.21989977e-01 -6.02298021e-01
-1.25514269e-01 -1.04789309e-01 -4.07252312e-01 -3.51894982e-02
1.48338035e-01 -1.60234487e+00 3.19413245e-01 5.72298050e-01
-1.21038824e-01 4.28543746e-01 1.07468891e+00 -7.79629052e-01
-1.16458654e+00 -5.12534261e-01 3.62191170e-01 -8.20689738e-01
4.45573866e-01 -1.56257540e-01 -4.09246564e-01 1.55081522e+00
2.83351302e-01 -6.64094090e-01 4.66110408e-01 2.85000771e-01
2.64591761e-02 5.83822072e-01 -1.01158047e+00 7.96249270e-01
1.12634039e+00 -2.67745674e-01 -9.99873340e-01 4.79822785e-01
7.79698849e-01 -6.40634000e-01 -7.43753374e-01 4.00509298e-01
6.05489016e-01 -1.35675836e+00 9.93700147e-01 -6.05858028e-01
3.34569395e-01 -3.03697348e-01 -2.87984192e-01 -1.26162100e+00
-7.66166985e-01 -7.26568520e-01 5.68357594e-02 5.79159260e-01
1.27000839e-01 -5.74557662e-01 7.24267066e-01 9.29016232e-01
8.58767927e-02 -4.54858392e-01 -8.70359838e-01 -4.35235918e-01
8.58604908e-02 -4.90778267e-01 3.78629148e-01 3.11097264e-01
5.06261468e-01 5.29592931e-01 -4.76577163e-01 3.38535547e-01
7.54411995e-01 -1.23678185e-01 7.54995823e-01 -1.22890532e+00
-8.55301380e-01 -1.54050544e-01 -1.44957319e-01 -1.14222181e+00
3.32744032e-01 -6.13387883e-01 5.78696072e-01 -1.79943514e+00
5.32494307e-01 -4.14026827e-01 1.32501900e-01 5.82224369e-01
7.88189396e-02 2.37881988e-02 3.33177209e-01 -2.36685686e-02
-8.96426260e-01 5.76994359e-01 1.24321687e+00 2.28901967e-01
-3.19745660e-01 3.04155320e-01 -1.01562536e+00 1.21213305e+00
6.04790270e-01 -5.46242952e-01 -5.58175027e-01 -2.25598186e-01
3.16146523e-01 6.65597856e-01 2.64107049e-01 -7.71914542e-01
3.91635835e-01 -5.25238156e-01 3.10861498e-01 -2.47858852e-01
6.18897200e-01 -5.36029220e-01 2.95143992e-01 6.42366827e-01
-7.38482058e-01 1.96868211e-01 -1.12162247e-01 8.01833928e-01
4.17274565e-01 -4.21461254e-01 6.18768096e-01 -6.30633831e-01
-5.17477095e-01 -1.22673817e-01 -9.42817032e-01 8.47226605e-02
8.42612028e-01 -1.30416518e-02 4.40440662e-02 -1.02609992e+00
-1.01946914e+00 3.78404438e-01 4.69324917e-01 -2.60971427e-01
3.72743666e-01 -9.16274488e-01 -4.52116877e-01 -8.96844491e-02
-3.18783939e-01 8.66541117e-02 4.25308049e-02 6.50407791e-01
-3.78362596e-01 5.95091403e-01 -1.11348897e-01 -1.33334264e-01
-9.38887417e-01 5.35370588e-01 4.48821187e-01 -7.68787622e-01
-7.65163362e-01 4.75799114e-01 5.05881846e-01 -5.74979842e-01
1.40340984e-01 -7.12975040e-02 -3.42640460e-01 -3.24140459e-01
4.03434664e-01 4.73758847e-01 -3.29602569e-01 -6.65890336e-01
-1.22220270e-01 1.51753157e-01 4.54772413e-02 -9.39719617e-01
1.49213827e+00 -2.84746468e-01 -2.09132154e-02 3.68481547e-01
5.06303981e-02 -1.97188750e-01 -1.64471781e+00 -1.18636727e-01
2.09448755e-01 -1.72668591e-01 7.37658963e-02 -7.41794705e-01
-2.82260984e-01 5.69067657e-01 3.15155764e-03 3.07595767e-02
7.90918648e-01 1.57601669e-01 1.83574662e-01 6.46727145e-01
1.19836509e+00 -9.92031813e-01 -1.46421585e-02 8.42992485e-01
8.32575679e-01 -7.06128955e-01 3.97361159e-01 -2.71333903e-01
-9.12570536e-01 7.04807758e-01 4.04745996e-01 -2.46295556e-01
2.81126410e-01 2.87584752e-01 -3.33319694e-01 -2.69804716e-01
-1.11309445e+00 -5.66987276e-01 -2.81408429e-01 6.71148360e-01
8.91925916e-02 2.14285702e-01 8.16691946e-03 5.55989444e-01
-6.82990372e-01 1.64872687e-02 9.69996274e-01 9.83681202e-01
-6.09994173e-01 -1.11613727e+00 -4.96331036e-01 2.43384257e-01
-1.82423621e-01 1.11861356e-01 -3.84754717e-01 9.74411309e-01
-2.15822294e-01 1.13230014e+00 -1.55944720e-01 -1.86893791e-01
8.58197734e-02 1.71253588e-02 1.05793583e+00 -8.77605855e-01
-2.07455307e-01 3.73185039e-01 1.20762490e-01 -8.88818383e-01
-3.84858221e-01 -8.54607642e-01 -1.38707078e+00 -2.17579141e-01
7.31680468e-02 2.73900807e-01 6.10534521e-03 1.32776523e+00
-1.05187275e-01 5.27877808e-01 1.84326202e-01 -1.29595983e+00
-7.02167392e-01 -8.68020952e-01 -7.57946908e-01 -4.71476512e-03
2.76188463e-01 -9.98363376e-01 -3.67001235e-01 -1.82651296e-01]
|
[4.003285884857178, 1.4393506050109863]
|
58369241-3c1d-43e6-8bf8-3a07a5e96dce
|
image-forgery-detection-based-on-the-fusion
|
1311.6934
| null |
http://arxiv.org/abs/1311.6934v1
|
http://arxiv.org/pdf/1311.6934v1.pdf
|
Image forgery detection based on the fusion of machine learning and block-matching methods
|
Dense local descriptors and machine learning have been used with success in
several applications, like classification of textures, steganalysis, and
forgery detection. We develop a new image forgery detector building upon some
descriptors recently proposed in the steganalysis field suitably merging some
of such descriptors, and optimizing a SVM classifier on the available training
set. Despite the very good performance, very small forgeries are hardly ever
detected because they contribute very little to the descriptors. Therefore we
also develop a simple, but extremely specific, copy-move detector based on
region matching and fuse decisions so as to reduce the missing detection rate.
Overall results appear to be extremely encouraging.
|
['Luisa Verdoliva', 'Diego Gragnaniello', 'Davide Cozzolino']
|
2013-11-27
| null | null | null | null |
['steganalysis']
|
['computer-vision']
|
[ 5.66310525e-01 -2.23696113e-01 4.32350766e-03 -1.63584054e-01
-6.06634080e-01 -1.73034415e-01 8.86051059e-01 1.54187799e-01
-3.23649108e-01 5.93534946e-01 -1.11576036e-01 -1.98450759e-01
1.98658362e-01 -8.22679102e-01 -3.43379706e-01 -1.06078196e+00
-2.47673005e-01 -4.28208485e-02 6.44442260e-01 -4.15471286e-01
7.90157914e-01 6.63002729e-01 -1.52314687e+00 4.81250852e-01
5.33420444e-01 1.09607422e+00 4.62938733e-02 8.05301189e-01
1.51553437e-01 8.06319237e-01 -7.23745525e-01 -4.22510892e-01
1.97358489e-01 -6.41215444e-01 -6.25478804e-01 5.01045942e-01
1.92410797e-02 -3.05927575e-01 -4.39703971e-01 1.12025726e+00
3.53963047e-01 -1.29463866e-01 7.13611722e-01 -7.64224172e-01
-4.21433300e-01 1.91466540e-01 -4.95444953e-01 4.31716323e-01
5.34225106e-01 -1.67986557e-01 4.46745098e-01 -9.31549430e-01
6.02939010e-01 8.93093050e-01 7.76406407e-01 1.09152153e-01
-8.63426268e-01 -5.22537589e-01 -6.60425305e-01 4.69055146e-01
-1.56011093e+00 -6.73664331e-01 8.47664118e-01 1.33219762e-02
8.43138158e-01 4.68795657e-01 4.53079790e-01 9.08866823e-01
6.02362096e-01 6.72844410e-01 1.34135604e+00 -8.94426644e-01
3.88146304e-02 3.81734043e-01 -4.64630932e-01 1.14506912e+00
1.13706172e-01 3.77086818e-01 -2.36570537e-01 -3.61798018e-01
7.52757728e-01 -7.03790132e-03 -2.28493407e-01 -4.87228394e-01
-1.00307322e+00 1.20834911e+00 2.10815638e-01 9.19738412e-01
-1.21192828e-01 2.35242732e-02 4.03400272e-01 9.61522281e-01
5.08982241e-01 3.24594021e-01 5.52460477e-02 5.73311448e-02
-1.03029621e+00 6.39814436e-02 8.36619496e-01 5.54145813e-01
7.23562837e-01 1.49003088e-01 3.01286519e-01 7.35593855e-01
9.34450421e-03 1.62725240e-01 6.25123382e-01 -5.53895414e-01
2.60836750e-01 2.51469582e-01 -3.28837074e-02 -1.67398703e+00
-1.76534712e-01 -2.36617744e-01 -8.04352760e-01 4.53538626e-01
3.33148181e-01 3.93826097e-01 -8.57743561e-01 8.97586823e-01
1.02223754e-01 -3.43707018e-02 -1.45618066e-01 4.83509868e-01
1.66018799e-01 5.02708077e-01 -2.95382559e-01 -6.78768530e-02
1.03359103e+00 -6.39870048e-01 -5.10419428e-01 4.04133536e-02
9.36470807e-01 -1.10261774e+00 2.10987598e-01 7.11246848e-01
-9.21401143e-01 -4.68626499e-01 -1.21317434e+00 3.73870760e-01
-6.18841648e-01 -9.34357345e-02 5.67074299e-01 1.34167624e+00
-1.12482524e+00 9.36441243e-01 -5.85036755e-01 -3.49178076e-01
3.10738236e-01 4.81012523e-01 -6.00237846e-01 -3.36123139e-01
-1.10683489e+00 9.63715792e-01 6.17877543e-01 -1.16882198e-01
-6.35389388e-01 2.10750088e-01 -8.79447103e-01 -1.34314984e-01
2.67884493e-01 -1.84312090e-02 4.84188586e-01 -1.04700637e+00
-1.43198681e+00 1.16275334e+00 1.46716356e-01 -5.99310517e-01
6.37347162e-01 5.94958663e-01 -8.88812840e-01 3.48244101e-01
-9.70393792e-02 2.13815331e-01 1.47551179e+00 -9.59168553e-01
-6.78327143e-01 -3.45816046e-01 -4.33289260e-01 -1.62851125e-01
-1.77119017e-01 3.32180023e-01 -2.37464994e-01 -1.00940919e+00
2.20580742e-01 -6.20882034e-01 -1.91793203e-01 -2.08321549e-02
-1.02018707e-01 -5.74547052e-02 1.32980883e+00 -8.98666084e-01
1.03507757e+00 -2.29441595e+00 -7.68336207e-02 7.01833546e-01
2.77684212e-01 5.83116472e-01 -8.85661617e-02 5.39170146e-01
-9.43332911e-02 -9.86316130e-02 -4.61796910e-01 -5.98571636e-02
-4.37312454e-01 2.25099355e-01 -1.26604721e-01 1.11248398e+00
2.17779592e-01 5.95165014e-01 -7.93313324e-01 -5.47628999e-01
7.05210090e-01 2.67616570e-01 -1.76616639e-01 -2.34892040e-01
4.62667465e-01 8.29822347e-02 -5.33835769e-01 8.66850793e-01
8.49687994e-01 -3.11104711e-02 9.09139812e-02 2.57290125e-01
1.01603322e-01 -2.18382597e-01 -1.21497214e+00 1.23819673e+00
-2.72634089e-01 7.78578639e-01 1.92630887e-01 -1.67590117e+00
1.24815488e+00 2.75064588e-01 5.12256861e-01 -7.34748125e-01
4.16150182e-01 4.76695895e-01 -1.98392227e-01 -5.86539209e-01
5.94610274e-01 -2.59753793e-01 -1.28942728e-01 2.81196892e-01
1.67959705e-01 -2.25027613e-02 -1.86506182e-01 -6.28554225e-02
1.41195476e+00 -2.42463663e-01 6.47640109e-01 -1.32260174e-01
9.08738971e-01 -8.08025226e-02 1.17758922e-02 8.65844607e-01
-4.00909007e-01 7.38336027e-01 1.96331009e-01 -4.15374845e-01
-1.24519932e+00 -4.73165989e-01 -1.96603596e-01 8.50026488e-01
3.75407487e-01 -1.42145202e-01 -5.58982134e-01 -8.51780295e-01
1.44849159e-02 1.24578536e-01 -4.73737597e-01 -2.70847827e-01
-7.35902548e-01 -7.54286766e-01 7.18387008e-01 5.65249063e-02
6.30867481e-01 -1.02700293e+00 -4.10713017e-01 4.14330512e-01
9.76850167e-02 -1.04870605e+00 -7.21807927e-02 1.92086712e-01
-9.15628731e-01 -1.08659458e+00 -8.59280467e-01 -1.13391078e+00
4.57879603e-01 5.60201108e-01 5.96473694e-01 4.66393709e-01
-7.32589543e-01 1.46069586e-01 -7.50981331e-01 8.33024010e-02
-9.81538832e-01 -2.55414665e-01 -1.79475814e-01 1.00119345e-01
2.68888831e-01 -2.18734175e-01 -2.33865499e-01 5.38329244e-01
-1.06834650e+00 -7.32338905e-01 6.25544667e-01 1.01740003e+00
1.31378785e-01 4.67652917e-01 2.44799644e-01 -7.58956969e-01
4.15474236e-01 -4.22550440e-01 -3.96240234e-01 8.48298296e-02
-4.87351149e-01 3.36715132e-02 7.57277966e-01 -2.26941317e-01
-6.47484958e-01 -6.61010891e-02 -4.32676524e-01 -3.71406794e-01
-2.93887466e-01 9.18375403e-02 7.42186829e-02 -1.15161848e+00
7.25661278e-01 5.97889304e-01 3.12587559e-01 -4.88237143e-01
2.01044194e-02 9.24287617e-01 4.80799049e-01 1.37368634e-01
1.00354278e+00 5.92949212e-01 1.01536766e-01 -1.29878020e+00
1.02494456e-01 -8.94107461e-01 -3.16437066e-01 5.25314547e-02
3.35259587e-01 -6.79551125e-01 -3.11319977e-01 8.02560985e-01
-9.13687170e-01 2.68320501e-01 -1.81179285e-01 2.79293805e-01
-6.22934401e-01 1.24005473e+00 -4.54727024e-01 -8.44675243e-01
-8.20403732e-03 -1.00155008e+00 8.32234263e-01 -2.29410052e-01
6.83279783e-02 -1.22750187e+00 -3.92827354e-02 1.41008228e-01
6.58326566e-01 5.72879970e-01 5.00366032e-01 -7.05898464e-01
-3.36400628e-01 -9.35122728e-01 -2.42067739e-01 6.03869438e-01
2.44043782e-01 -3.98890704e-01 -6.51859760e-01 -5.95361173e-01
4.52085704e-01 -3.21456552e-01 1.18827176e+00 -1.59652084e-02
1.16507876e+00 -1.92633942e-01 -5.84730089e-01 6.47446632e-01
1.58766222e+00 2.44185477e-01 9.71635520e-01 3.24858338e-01
2.84490228e-01 5.21851897e-01 4.73532349e-01 4.53638315e-01
-2.40486890e-01 8.39626849e-01 3.68221402e-01 -9.11457464e-02
-1.70930132e-01 1.21452041e-01 3.47609997e-01 5.48338711e-01
-1.45440191e-01 -3.83976161e-01 -5.05343914e-01 5.90624869e-01
-1.44524682e+00 -1.23851478e+00 -1.60801604e-01 2.12225938e+00
3.96644503e-01 1.58139482e-01 4.31892388e-02 6.07042015e-01
9.17838514e-01 4.93765801e-01 1.83470726e-01 -6.64767742e-01
-3.77422839e-01 4.42850292e-01 9.08336997e-01 5.26306152e-01
-1.46941280e+00 8.87024343e-01 7.28484344e+00 1.39595544e+00
-1.02394927e+00 2.07548693e-01 4.37518954e-01 7.26575851e-01
7.13299066e-02 4.07892279e-02 -4.61257517e-01 5.53404033e-01
6.74972832e-01 4.13918048e-01 2.32194588e-01 6.59430504e-01
-7.78613687e-02 -4.10639912e-01 -3.79658610e-01 1.21198738e+00
6.30889595e-01 -1.31475639e+00 -9.68171209e-02 3.85917425e-01
6.52625322e-01 -3.07963192e-01 7.91889674e-04 -7.04161152e-02
-2.48716176e-01 -8.97065282e-01 3.26195180e-01 3.37870181e-01
8.24699998e-01 -7.90384650e-01 8.91351581e-01 2.91349411e-01
-9.09499109e-01 -1.43204436e-01 -7.23034859e-01 9.16188583e-02
-4.46458012e-02 6.73262835e-01 -8.24320734e-01 5.21961331e-01
3.01881343e-01 5.38349986e-01 -5.03472745e-01 1.19959521e+00
1.40396520e-01 4.85491306e-01 -4.95218098e-01 -2.92074438e-02
4.88128424e-01 4.26698290e-03 7.57139444e-01 1.36790812e+00
4.83972698e-01 -1.23056009e-01 3.44498865e-02 5.16650021e-01
2.29726270e-01 2.61664361e-01 -9.24184501e-01 2.23238856e-01
-4.25978377e-02 9.71411526e-01 -8.75560820e-01 -2.92516828e-01
-3.97175550e-01 1.49389029e+00 -2.38675490e-01 -1.61284715e-01
-5.46118975e-01 -8.50319803e-01 2.56432444e-01 1.44793883e-01
9.18905973e-01 -2.46529698e-01 1.25853075e-02 -1.24558055e+00
-1.41249910e-01 -8.95143390e-01 1.96782425e-01 -9.26000252e-02
-9.31638181e-01 3.62140179e-01 -1.87889412e-01 -1.32725155e+00
-5.94659507e-01 -7.78710961e-01 -6.64385557e-01 5.47109127e-01
-1.40585375e+00 -8.91848028e-01 -2.63267141e-02 8.78346801e-01
4.11765784e-01 -5.83318293e-01 6.65801048e-01 1.42617434e-01
-1.27781197e-01 7.34703898e-01 6.32996082e-01 1.58226728e-01
5.63215792e-01 -8.16940188e-01 4.33448911e-01 8.49349380e-01
1.67242408e-01 1.17772952e-01 8.82871568e-01 -5.95009327e-01
-1.11627698e+00 -7.07787037e-01 1.03491771e+00 -2.15942785e-01
5.92594504e-01 -2.31876910e-01 -9.21145320e-01 1.03673823e-01
-1.29951015e-01 2.79190570e-01 2.20793426e-01 -4.84213233e-01
-2.72818893e-01 5.89973070e-02 -1.50093818e+00 -1.12608761e-01
7.03879297e-01 -5.95083237e-01 -4.53036219e-01 4.49223191e-01
-1.42090753e-01 -1.70089751e-01 -6.52339339e-01 9.74351913e-02
5.01782179e-01 -1.34985185e+00 1.06006646e+00 -8.17762911e-02
7.84071237e-02 2.52900477e-02 -9.70497727e-02 -9.29149747e-01
-4.14150715e-01 -7.27243125e-01 -8.45780373e-02 8.17054808e-01
-1.17494583e-01 -7.41211653e-01 1.06497383e+00 -4.21584010e-01
8.11974853e-02 -3.63851666e-01 -1.33579743e+00 -1.08698869e+00
-2.26060063e-01 -7.66330585e-02 2.16202855e-01 1.06525934e+00
1.23255379e-01 -3.43455374e-01 -1.03133750e+00 -1.31174132e-01
9.18814540e-01 -6.63265511e-02 4.54333037e-01 -9.14930105e-01
-4.16346520e-01 -3.92066091e-01 -1.30088603e+00 -8.04906487e-01
-1.73376292e-01 -6.38928771e-01 -2.48157587e-02 -6.92157626e-01
-1.40193865e-01 -5.57864845e-01 -2.36141667e-01 3.95535707e-01
9.17730778e-02 8.97276163e-01 -1.80523038e-01 3.42470706e-01
-4.21084732e-01 1.94639087e-01 8.88841271e-01 -1.50784120e-01
1.76665157e-01 2.18831122e-01 -3.83265644e-01 5.59703171e-01
8.08474481e-01 -5.33659816e-01 2.86248535e-01 1.95004269e-01
-3.27019840e-01 2.15212613e-01 4.63324875e-01 -1.40978360e+00
-1.08749215e-02 1.66281939e-01 5.89143813e-01 -1.54363558e-01
2.63871968e-01 -8.57522011e-01 -7.07211122e-02 9.40748811e-01
7.99874812e-02 -2.39563435e-01 -2.66713858e-01 7.89376557e-01
-5.32376468e-01 -5.64067543e-01 1.09689224e+00 -3.11596155e-01
-1.12797987e+00 -5.95519207e-02 -4.82481182e-01 -5.52978575e-01
1.32705498e+00 -9.25767839e-01 -4.00565267e-02 -4.87835050e-01
-6.57598794e-01 -6.13045573e-01 6.83448434e-01 3.66331905e-01
7.89511502e-01 -1.08426023e+00 -7.49516487e-01 6.72616541e-01
2.76918590e-01 -9.56017673e-01 1.58668607e-02 7.87799656e-01
-9.11722004e-01 4.72074687e-01 -3.95517290e-01 -3.92328888e-01
-1.44890404e+00 9.02520895e-01 1.50361702e-01 -4.21542108e-01
-7.21265435e-01 7.53342509e-01 -4.07916605e-01 1.77175209e-01
-4.71072271e-02 7.34044984e-02 -1.25712320e-01 -1.16651841e-01
6.22102916e-01 6.38661802e-01 3.22644025e-01 -8.69448662e-01
-3.79030704e-01 6.86882257e-01 -1.96821213e-01 1.95494443e-01
1.01751482e+00 -2.89627880e-01 -7.67772198e-02 -2.45708346e-01
1.60256362e+00 1.07403368e-01 -8.73052657e-01 -2.82618433e-01
6.83995709e-02 -1.01031554e+00 2.10331485e-01 -1.45059109e-01
-9.02495503e-01 7.38696933e-01 8.99907410e-01 7.00835586e-01
1.16048038e+00 6.92928582e-03 8.33889663e-01 2.94159621e-01
7.35147715e-01 -1.09934139e+00 7.36378133e-02 2.98705757e-01
5.30618668e-01 -1.32200849e+00 2.47252673e-01 -4.80138063e-01
-3.02949280e-01 1.32648933e+00 -3.15734506e-01 -7.02678323e-01
6.70869648e-01 2.49767244e-01 -2.80582130e-01 -6.38669878e-02
-1.58833832e-01 -7.89712667e-02 -2.26930529e-02 9.28335369e-01
-5.79504818e-02 -1.43180236e-01 -6.00317776e-01 -6.60312951e-01
2.64970779e-01 -2.64877323e-02 4.42708731e-01 1.19607329e+00
-9.53057051e-01 -1.38996625e+00 -8.07486117e-01 4.11027938e-01
-7.97749043e-01 6.02831133e-02 -2.71981746e-01 9.05327559e-01
2.40991265e-01 8.55352104e-01 -2.04420924e-01 -6.15736961e-01
-1.12179860e-01 -5.66669293e-02 6.24466062e-01 -1.31567135e-01
-5.48450232e-01 6.24992773e-02 -4.37208787e-02 -7.12318838e-01
-4.43010181e-01 -6.55828297e-01 -3.68365645e-01 -6.02166355e-01
-6.69778287e-01 6.37644976e-02 7.11855114e-01 1.20885193e+00
-1.31872803e-01 -1.11823037e-01 1.19992185e+00 -9.26450431e-01
-7.20391929e-01 -7.66636848e-01 -1.09302914e+00 4.12539452e-01
5.25569916e-01 -4.16315347e-01 -5.40144622e-01 -1.05137207e-01]
|
[4.331357479095459, 8.033811569213867]
|
1bf3efac-16ee-4635-abf4-717c4141ebc6
|
cross-lingual-prosody-transfer-for-expressive
|
2306.11658
| null |
https://arxiv.org/abs/2306.11658v1
|
https://arxiv.org/pdf/2306.11658v1.pdf
|
Cross-lingual Prosody Transfer for Expressive Machine Dubbing
|
Prosody transfer is well-studied in the context of expressive speech synthesis. Cross-lingual prosody transfer, however, is challenging and has been under-explored to date. In this paper, we present a novel solution to learn prosody representations that are transferable across languages and speakers for machine dubbing of expressive multimedia contents. Multimedia contents often contain field recordings. To enable prosody transfer from noisy audios, we introduce a novel noise modelling module that disentangles noise conditioning from prosody conditioning, and thereby gains independent control of noise levels in the synthesised speech. We augment noisy training data with clean data to improve the ability of the model to map the denoised reference audio to clean speech. Our proposed system can generate speech with context-matching prosody and closes the gap between a strong baseline and human expressive dialogs by 11.2%.
|
['Vincent Pollet', 'Ravichander Vipperla', 'Patrick Lumban Tobing', 'Mikolaj Babianski', 'Duo Wang', 'Jakub Swiatkowski']
|
2023-06-20
| null | null | null | null |
['expressive-speech-synthesis', 'speech-synthesis']
|
['speech', 'speech']
|
[ 2.00470254e-01 1.10293940e-01 -2.71541979e-02 -2.49113962e-01
-1.60266280e+00 -7.00975895e-01 4.28344369e-01 -2.96363860e-01
-1.65330172e-01 7.13047624e-01 8.54361236e-01 1.19160734e-01
4.39777166e-01 -5.48730254e-01 -8.37153435e-01 -5.63653231e-01
4.25473630e-01 1.89369202e-01 9.34515893e-02 -5.09018302e-01
-5.28251231e-01 2.45090038e-01 -1.55125558e+00 7.84912348e-01
5.65361440e-01 9.05145526e-01 4.82450664e-01 9.60344732e-01
-3.53383154e-01 8.37240279e-01 -1.11907387e+00 -1.07772134e-01
1.33915424e-01 -7.57555783e-01 -4.41146672e-01 -2.57393837e-01
4.14836377e-01 -2.71467447e-01 -3.34560066e-01 9.14675713e-01
9.04580355e-01 2.30723321e-01 5.46660841e-01 -1.00456715e+00
-3.77876848e-01 1.22738194e+00 7.75706442e-03 -2.43227417e-03
4.18711692e-01 1.57805741e-01 9.98981476e-01 -7.29565144e-01
5.65739334e-01 1.54217887e+00 5.33235073e-01 1.15268517e+00
-1.62380064e+00 -1.26601696e+00 -1.65631711e-01 -1.45201996e-01
-1.25024498e+00 -9.84505236e-01 1.07204819e+00 -3.02154928e-01
7.07919002e-01 4.68614310e-01 3.79849404e-01 1.76998639e+00
-3.71058673e-01 8.19433033e-01 9.69870806e-01 -5.27467072e-01
6.94735572e-02 3.28689665e-01 -3.71313065e-01 -1.25904486e-01
-7.02010989e-01 3.23605686e-01 -1.13668990e+00 5.63673228e-02
5.44771254e-01 -7.84456432e-01 -4.18990999e-01 1.05237000e-01
-9.59217370e-01 6.49955034e-01 2.07594618e-01 4.24813956e-01
-2.93973297e-01 2.80434281e-01 7.05855548e-01 4.40621167e-01
3.44330102e-01 4.02640134e-01 -2.01776862e-01 -2.87400156e-01
-1.07571149e+00 2.80901283e-01 8.66921544e-01 9.84192848e-01
3.02284032e-01 6.95623457e-01 -2.97761619e-01 1.36456633e+00
1.13059096e-01 7.21922040e-01 4.85303104e-01 -9.66451466e-01
5.40178061e-01 -2.02801332e-01 4.55005281e-02 -3.99424374e-01
-2.21496187e-02 -1.32061929e-01 -5.92858434e-01 1.47456944e-01
2.39492938e-01 -5.54602087e-01 -6.00946009e-01 2.18422055e+00
1.64137796e-01 2.80602604e-01 2.84558088e-01 9.01724577e-01
8.69016171e-01 1.18029749e+00 3.40358973e-01 -3.68037909e-01
1.23819113e+00 -9.91043150e-01 -1.31536651e+00 -1.67387635e-01
-6.73236102e-02 -1.04649591e+00 1.54419208e+00 4.79739815e-01
-1.47482073e+00 -9.68501031e-01 -1.22652340e+00 -2.99408942e-01
-3.80769558e-02 -1.01421885e-01 -8.12610835e-02 7.27601290e-01
-8.34473670e-01 2.70794928e-01 -6.53073251e-01 9.78107005e-02
-6.40000403e-02 2.74547666e-01 -2.57193327e-01 4.33363825e-01
-1.45873678e+00 5.87224305e-01 3.97312135e-01 -4.65716332e-01
-1.05565274e+00 -1.09767568e+00 -9.46713030e-01 2.56095618e-01
4.42048162e-01 -2.97458857e-01 1.74657440e+00 -1.19419611e+00
-2.24948382e+00 5.53858340e-01 4.42016609e-02 -5.96441984e-01
4.03162986e-01 -6.86012983e-01 -6.94694579e-01 1.52399629e-01
-1.42036960e-01 8.73581707e-01 1.00051248e+00 -1.45228875e+00
-4.67729807e-01 1.88311502e-01 -5.12581527e-01 3.98269385e-01
-3.75512332e-01 3.59227240e-01 -4.33781087e-01 -1.09382570e+00
-4.78440881e-01 -7.61278927e-01 1.88460663e-01 -1.64407954e-01
-2.52848148e-01 1.26534298e-01 7.54819214e-01 -7.97205150e-01
1.17815113e+00 -2.41566753e+00 3.07912916e-01 -1.90633833e-02
-4.10948157e-01 2.99050808e-01 -2.68942505e-01 4.40871835e-01
-1.40327007e-01 -1.48228249e-02 -8.61706287e-02 -6.50144875e-01
2.57456154e-01 1.95307761e-01 -8.20127845e-01 -1.09694742e-01
2.99381018e-01 5.75492382e-01 -6.49007618e-01 -3.83412927e-01
1.47742108e-01 8.10109019e-01 -6.24762118e-01 7.40369022e-01
-4.86624926e-01 7.00680852e-01 8.69114231e-03 2.58996785e-01
4.00223553e-01 7.06180274e-01 2.41737664e-01 -2.44103685e-01
-1.61291495e-01 8.05886626e-01 -1.22209334e+00 1.95605588e+00
-8.39842379e-01 5.24347067e-01 8.86869967e-01 -4.53683138e-01
1.06333017e+00 9.17829454e-01 2.82412082e-01 -5.57348311e-01
1.75371259e-01 1.26558423e-01 -5.69052398e-02 -3.02291721e-01
4.54090267e-01 -6.62942290e-01 -2.00543761e-01 1.24142684e-01
4.69313204e-01 -8.16131175e-01 -1.77173138e-01 -3.58912289e-01
5.88182271e-01 1.03541747e-01 2.37247452e-01 -1.89059108e-01
6.76213861e-01 -3.99803191e-01 4.43589807e-01 5.17786384e-01
-1.35196432e-01 7.78713763e-01 2.33323053e-01 3.44342798e-01
-9.05003428e-01 -1.47764730e+00 1.00795791e-01 1.68547165e+00
-3.16745728e-01 -4.39618558e-01 -1.22755706e+00 -2.46074349e-01
-2.76539057e-01 1.09474480e+00 -1.22581102e-01 -8.00147578e-02
-8.95024538e-01 -5.02944030e-02 1.09236264e+00 5.07528245e-01
1.73541740e-01 -1.42096245e+00 1.49412274e-01 4.20217216e-01
-4.52090681e-01 -1.22387385e+00 -9.82608020e-01 3.60232323e-01
-2.76912093e-01 -4.20734346e-01 -6.73884332e-01 -1.05711997e+00
-2.55705267e-01 -2.08096281e-01 1.06498361e+00 -4.87544596e-01
3.79103310e-02 3.03101301e-01 -1.50213018e-01 -6.27690494e-01
-1.39698267e+00 8.04507807e-02 3.50196451e-01 -9.31312982e-03
7.98063949e-02 -7.25281298e-01 -1.34170115e-01 3.48492593e-01
-1.00520778e+00 -2.98652425e-02 2.80567795e-01 9.36532557e-01
7.53461778e-01 -2.89223313e-01 1.07727516e+00 -4.86570150e-01
9.69024241e-01 -2.66371161e-01 -5.73902845e-01 -1.00776196e-01
8.64224136e-02 -2.05486435e-02 8.42631459e-01 -8.53657186e-01
-1.49643302e+00 7.71591291e-02 -7.62907028e-01 -6.26144171e-01
-3.80647987e-01 1.68322220e-01 -7.80580878e-01 4.23041552e-01
9.70302403e-01 -6.40671700e-02 -1.30102299e-02 -7.29602873e-01
7.63082922e-01 1.07811570e+00 1.07044220e+00 -8.25737476e-01
6.12937748e-01 -3.54020204e-03 -4.86704201e-01 -1.19885457e+00
-5.42500317e-01 -3.16800177e-01 -1.34358019e-01 -8.19394644e-03
9.37895894e-01 -1.27523661e+00 -5.92386007e-01 2.38679960e-01
-1.15823734e+00 -6.57765150e-01 -6.69050395e-01 5.63587308e-01
-1.04681289e+00 6.21207058e-02 -9.91406441e-01 -9.63304937e-01
-5.67837119e-01 -1.21826673e+00 1.06599879e+00 -1.18184537e-01
-6.97569668e-01 -5.11838078e-01 1.96897328e-01 4.31706488e-01
4.49662626e-01 -7.83227310e-02 7.35732794e-01 -6.59542799e-01
-2.43453845e-01 1.98035672e-01 3.19298297e-01 8.82588387e-01
3.11359525e-01 -2.30362952e-01 -1.54448843e+00 -5.96532458e-03
3.23190331e-01 -5.92866480e-01 5.23437142e-01 3.42176974e-01
8.26331675e-01 -3.87689501e-01 2.00763956e-01 4.93936688e-01
6.82813466e-01 2.00186670e-01 5.52346587e-01 -1.57430753e-01
2.92824209e-01 7.65484750e-01 4.68118429e-01 1.52718648e-01
-1.40355751e-01 9.82681811e-01 -7.86126330e-02 -4.78641763e-02
-8.89247596e-01 -6.28576100e-01 8.23737383e-01 1.26203334e+00
3.68387848e-01 -3.12815607e-01 -4.93548512e-01 6.17961884e-01
-1.35185564e+00 -9.49264884e-01 3.58951002e-01 1.87204099e+00
1.53256238e+00 1.35672897e-01 2.10528061e-01 1.65968642e-01
7.32368410e-01 1.69594705e-01 -2.59066403e-01 -5.29645443e-01
-1.57199875e-01 6.62422895e-01 -9.03149787e-03 9.61941898e-01
-9.98566866e-01 1.22552729e+00 6.30865049e+00 1.26581800e+00
-1.17205465e+00 1.85273483e-01 2.03864917e-01 -3.27751637e-01
-2.85596222e-01 -5.75487077e-01 -7.20089495e-01 2.59319335e-01
1.29333889e+00 -2.22438380e-01 6.89369202e-01 9.88695145e-01
4.44449008e-01 4.88289684e-01 -1.42315578e+00 1.08962059e+00
-1.64585933e-03 -1.10936236e+00 9.49625578e-03 -4.54374969e-01
6.95064723e-01 -1.66516662e-01 4.17814493e-01 5.98051250e-01
1.48385003e-01 -1.14363611e+00 9.96262014e-01 2.68013895e-01
9.77185249e-01 -8.27665150e-01 3.21332216e-01 3.65746915e-01
-1.10984826e+00 6.06722981e-02 1.71790630e-01 2.18583792e-01
4.16088820e-01 -6.97405934e-02 -1.16187656e+00 1.59829378e-01
5.38434744e-01 8.86022672e-02 2.10307404e-01 4.21745509e-01
-3.54582012e-01 1.07534468e+00 -3.10425997e-01 2.45238006e-01
-1.50936186e-01 3.98188494e-02 7.94343948e-01 1.59659934e+00
3.95922631e-01 -1.00841455e-01 1.49558425e-01 1.09615922e+00
-4.45473313e-01 4.33417141e-01 -5.38990319e-01 -1.82567209e-01
4.84896183e-01 7.68845916e-01 1.41028445e-02 -1.93814158e-01
-2.95193702e-01 8.60355556e-01 2.88499956e-04 3.81995261e-01
-7.27139413e-01 -4.83615622e-02 1.14884782e+00 8.94549713e-02
1.13229431e-01 1.41667174e-02 -2.58832723e-01 -8.10390472e-01
-2.30596378e-01 -1.32121193e+00 9.23915282e-02 -8.45568180e-01
-1.40713418e+00 7.98157573e-01 -2.10345164e-02 -1.00935125e+00
-7.83090115e-01 -3.97554129e-01 -4.79642659e-01 1.16340196e+00
-1.28964400e+00 -1.02836621e+00 7.47881085e-02 6.16135955e-01
1.08671987e+00 -3.79117936e-01 1.21277595e+00 4.45814848e-01
-7.18846545e-02 7.16162145e-01 -5.44230752e-02 1.17407136e-01
1.10113192e+00 -1.18483710e+00 3.95352393e-01 5.90391695e-01
1.48504823e-01 4.80115891e-01 1.07347965e+00 -4.04278904e-01
-8.64303231e-01 -1.12898362e+00 5.37957489e-01 -2.26394862e-01
7.73416638e-01 -8.71635854e-01 -1.16275430e+00 3.64122510e-01
6.94823980e-01 -1.93464205e-01 9.44435179e-01 -2.19754621e-01
-3.75187069e-01 -4.67538267e-01 -9.69870329e-01 9.66708243e-01
6.49264872e-01 -1.03376865e+00 -7.05991030e-01 -1.01061523e-01
1.24779117e+00 -4.54982519e-01 -7.95471728e-01 2.19032630e-01
3.32400471e-01 -7.72174895e-01 1.10259938e+00 -5.16311586e-01
1.06849022e-01 -1.09978043e-01 -5.19982576e-01 -1.47524631e+00
2.38114819e-01 -1.23889863e+00 -3.94230522e-02 1.75873542e+00
4.91566539e-01 4.44785617e-02 4.67758089e-01 3.10109109e-01
-3.46263677e-01 1.35188177e-01 -8.96192133e-01 -7.97038436e-01
3.98726285e-01 -7.50454545e-01 5.83747804e-01 5.36617875e-01
1.42199367e-01 6.57953560e-01 -6.86092377e-01 1.68864056e-01
1.60890207e-01 -2.25027651e-01 9.17925298e-01 -6.76962554e-01
-5.68061709e-01 -4.16063011e-01 1.19144939e-01 -1.08003914e+00
5.14007032e-01 -7.51192153e-01 5.53831160e-01 -8.84159386e-01
-4.65128034e-01 -6.46806136e-02 -1.81246400e-01 1.54033527e-01
-1.12190686e-01 9.00858268e-02 5.40610969e-01 -3.15591305e-01
4.08269688e-02 8.25481534e-01 1.05485272e+00 -2.17387661e-01
-6.24835491e-01 1.46978050e-01 -4.13525522e-01 7.43136823e-01
9.47124958e-01 -5.09497941e-01 -5.96837997e-01 -4.96455245e-02
-3.93819273e-01 5.47238111e-01 -2.15073135e-02 -1.11598527e+00
6.72450438e-02 -6.78350180e-02 -3.68013382e-02 -2.93241113e-01
9.97325301e-01 -9.20250833e-01 2.54160941e-01 2.58039050e-02
-9.12925780e-01 -4.13750291e-01 6.84956193e-01 4.54341054e-01
-6.06443226e-01 -1.82961866e-01 9.18154716e-01 3.18865106e-02
-2.17196494e-01 -2.41457552e-01 -5.22863507e-01 3.46852779e-01
5.79908192e-01 3.73025626e-01 7.43036624e-03 -7.82074690e-01
-9.88848686e-01 -2.18937546e-01 -2.43596435e-02 6.36299074e-01
4.04055744e-01 -1.39034653e+00 -7.34425426e-01 2.64796555e-01
-1.82431415e-01 -2.43636161e-01 3.61347824e-01 1.53297767e-01
-2.31976181e-01 1.16492815e-01 -7.26615489e-02 -3.84756625e-01
-1.43523812e+00 6.61978424e-01 4.56524938e-01 2.31157206e-02
-4.92298275e-01 6.99823260e-01 4.41846371e-01 -4.50879276e-01
8.01767290e-01 -4.00300980e-01 -1.87681373e-02 5.40356822e-02
7.02604890e-01 1.86615989e-01 -2.31595542e-02 -7.91042328e-01
6.38697743e-02 2.31483713e-01 2.78402835e-01 -7.35708654e-01
1.09676361e+00 -1.73419327e-01 2.59055167e-01 7.50097394e-01
1.21656430e+00 6.84181750e-01 -1.47630799e+00 -1.96038783e-01
-1.00510031e-01 1.07268773e-01 5.88791911e-03 -9.54654515e-01
-6.80462539e-01 9.92863774e-01 4.59757626e-01 2.13304147e-01
1.21193492e+00 -1.06982104e-01 1.03771484e+00 2.88518757e-01
-1.51479304e-01 -1.34787095e+00 2.47512519e-01 5.25330007e-01
1.38107336e+00 -9.33896661e-01 -7.36374736e-01 -6.12270117e-01
-1.05872500e+00 9.38897252e-01 5.74826241e-01 3.64746084e-03
5.57973146e-01 9.26441252e-01 5.98495126e-01 4.37751979e-01
-6.66480899e-01 -1.76134333e-01 4.84966815e-01 8.42610002e-01
4.77399915e-01 1.41457483e-01 1.80220962e-01 1.20704949e+00
-7.33859301e-01 -1.95918754e-01 1.41764045e-01 2.86145240e-01
-4.21009272e-01 -1.28747153e+00 -7.10618556e-01 -5.07332146e-01
-6.74410045e-01 -2.75418431e-01 -4.39332604e-01 6.67573929e-01
3.24200769e-03 1.22378397e+00 -7.33865872e-02 -1.62996426e-01
6.78374410e-01 6.93047583e-01 9.54900831e-02 -6.57827616e-01
-9.60717142e-01 8.34127307e-01 4.22726303e-01 -3.36092085e-01
-1.94209054e-01 -4.44115818e-01 -1.25324953e+00 1.09023251e-01
-9.29580629e-02 1.94211662e-01 6.01860583e-01 5.31299412e-01
4.50623892e-02 9.32389259e-01 4.67585236e-01 -9.52620506e-01
-7.98348188e-01 -1.12866247e+00 -5.74351311e-01 4.28203166e-01
5.42652428e-01 -2.44679332e-01 -2.78195053e-01 3.59236807e-01]
|
[15.017630577087402, 6.531139373779297]
|
2ef5abf3-d3ec-4ec9-a252-bf9d399389e6
|
investor-s-sentiment-in-multi-agent-model-of
|
1208.3083
| null |
http://arxiv.org/abs/1208.3083v4
|
http://arxiv.org/pdf/1208.3083v4.pdf
|
Investor's sentiment in multi-agent model of the continuous double auction
|
We introduce and treat rigorously a new multi-agent model of the continuous
double auction or in other words the order book (OB). It is designed to explain
collective behaviour of the market when new information affecting the market
arrives. The novel feature of the model is two additional slow changing
parameters, the so-called sentiment functions. These sentiment functions
measure the conception of the fair price of two groups of investors, namely,
bulls and bears. Our model specifies differential equations for the time
evolution of sentiment functions and constitutes a nonlinear Markov process
which exhibits long term correlations. We explain the intuition behind
equations for sentiment functions and present numerical simulations which show
that the behaviour of our model is similar to the behaviour of the real market.
We also obtain a diffusion limit of the model, the Ornstein-Uhlenbeck type
process with variable volatility. The volatility is proportional to the
difference of opinions of bulls and bears about the fair price of a security.
The paper is complimentary to our previous work where mathematical proofs are
presented.
|
[]
|
2016-02-17
| null | null | null | null |
['mathematical-proofs']
|
['miscellaneous']
|
[-5.54606676e-01 -5.60529344e-02 1.59668356e-01 7.33585656e-02
1.36160076e-01 -8.92627180e-01 6.45132840e-01 2.43480444e-01
-7.98465610e-01 9.25707102e-01 -1.04144119e-01 3.30368360e-03
-3.51969510e-01 -9.43778157e-01 -5.37198484e-01 -1.01909685e+00
-2.87430137e-01 7.41417587e-01 7.16148838e-02 -6.54893100e-01
3.75226200e-01 1.29136249e-01 -1.23198020e+00 -2.61954516e-01
9.70472842e-02 1.35696459e+00 -1.48942038e-01 9.06435132e-01
-1.72496200e-01 9.30296361e-01 -7.51949549e-01 -5.97597361e-01
7.02009499e-01 -4.02890474e-01 -3.29358101e-01 -1.03117630e-01
-8.15365374e-01 -1.85434759e-01 1.06182314e-01 1.20317280e+00
1.27975598e-01 -5.71491085e-02 8.41336966e-01 -1.36865294e+00
-1.00993919e+00 7.32174158e-01 -8.82219553e-01 4.88065034e-01
-1.57118306e-01 -2.83376217e-01 1.13357246e+00 -3.01257193e-01
3.12631220e-01 1.09582257e+00 4.63561833e-01 6.22170806e-01
-9.53535855e-01 -3.70939910e-01 2.19841108e-01 -1.64105728e-01
-9.76911902e-01 1.93307683e-01 3.55224609e-01 -4.58634377e-01
2.30061606e-01 3.73639822e-01 8.52426231e-01 5.77999711e-01
1.08090258e+00 5.20476699e-01 1.42162716e+00 -2.87683696e-01
7.05277741e-01 4.19134885e-01 3.35890621e-01 1.17891528e-01
6.77468061e-01 2.36874186e-02 -3.53947401e-01 -5.34450829e-01
6.66771948e-01 2.69280642e-01 -1.46723852e-01 -4.65186298e-01
-8.08532715e-01 1.01033604e+00 -1.69730961e-01 2.28083462e-01
-8.91187072e-01 2.12138996e-01 -3.57518494e-02 7.04262197e-01
7.12027550e-01 -8.66852179e-02 -7.33640432e-01 -1.22831672e-01
-2.15717599e-01 3.82128209e-01 1.38575554e+00 6.09681070e-01
2.69840091e-01 -2.32718080e-01 2.69128680e-01 1.16934054e-01
5.54479659e-01 9.28434074e-01 3.90504271e-01 -8.44149113e-01
5.80461286e-02 8.40112567e-02 6.23287320e-01 -6.47241533e-01
-3.80974174e-01 -4.64919448e-01 -6.32391334e-01 4.51414943e-01
5.47184706e-01 -7.12349653e-01 1.01411738e-01 1.53276515e+00
9.91426036e-02 -2.13476077e-01 3.62601310e-01 4.79125500e-01
2.43611820e-02 6.79734588e-01 -2.74835765e-01 -9.01411355e-01
1.25533044e+00 -5.40389061e-01 -9.83250022e-01 4.65710014e-01
6.52025938e-02 -6.72633410e-01 1.67165846e-01 2.99823135e-01
-1.42497301e+00 2.61942923e-01 -6.01607561e-01 1.02583814e+00
-4.25350130e-01 -8.63556504e-01 3.36417377e-01 6.92286074e-01
-1.14245141e+00 6.07265055e-01 -7.71556437e-01 -1.16082966e-01
-1.59450695e-01 4.03849244e-01 2.59826630e-01 7.68991768e-01
-9.87187803e-01 9.15210545e-01 -5.20151734e-01 1.77438483e-01
-5.59887946e-01 -6.29582345e-01 -2.32193246e-01 3.17379609e-02
3.01728338e-01 -7.55651116e-01 1.71776199e+00 -1.11222816e+00
-1.64552319e+00 6.08135819e-01 2.08687425e-01 -6.49787247e-01
8.96081150e-01 1.75593063e-01 -2.45129168e-01 -3.00707798e-02
1.33631736e-01 -2.88102865e-01 7.23661005e-01 -1.27758467e+00
-7.95745254e-01 -4.43905681e-01 -1.95811111e-02 1.98793024e-01
-3.88301015e-02 1.34383366e-01 5.76200783e-01 -6.69677019e-01
-3.61154616e-01 -9.81836557e-01 -2.86205679e-01 -4.41148043e-01
1.76571041e-01 -1.42831519e-01 4.96856838e-01 -7.26913065e-02
8.05198550e-01 -1.90709436e+00 -1.58786669e-01 2.40073249e-01
2.20359191e-01 -3.89185458e-01 3.25621217e-02 9.91158903e-01
1.70015078e-02 3.66281569e-01 7.13157430e-02 3.69003043e-02
3.23849410e-01 2.46086150e-01 -3.49570900e-01 8.80325973e-01
-1.27434209e-01 5.89001119e-01 -7.16670215e-01 3.35451961e-01
-5.12339473e-01 1.52382448e-01 -2.42286503e-01 -3.52857113e-01
-1.08991578e-01 1.69029057e-01 -7.86582589e-01 2.23702520e-01
7.87132740e-01 -1.94338128e-01 -1.04203686e-01 4.52671021e-01
-5.59709728e-01 -3.20100516e-01 -1.32698047e+00 3.30869913e-01
-2.55856961e-01 2.03765169e-01 4.87175107e-01 -5.70420325e-01
7.21100271e-01 4.47416127e-01 5.19579232e-01 -3.70776147e-01
4.89163369e-01 2.85429448e-01 2.07994372e-01 -3.53051238e-02
4.93852526e-01 -8.86732876e-01 -1.29675835e-01 1.07226384e+00
-3.44985455e-01 -2.53743269e-02 2.98762947e-01 2.09955677e-01
1.03598046e+00 -4.92448539e-01 3.51809978e-01 -8.54534924e-01
3.92042816e-01 -3.33704799e-01 5.23152888e-01 8.44145477e-01
-3.25856477e-01 -1.13695592e-01 1.01887894e+00 -5.77410817e-01
-9.70536530e-01 -1.01387715e+00 5.38332574e-02 5.27706683e-01
5.03263533e-01 8.26847851e-02 -6.65553570e-01 -1.39106736e-01
3.87950271e-01 5.75460613e-01 -9.82911885e-01 1.08071268e-01
-5.92992501e-03 -1.29975259e+00 -2.37572685e-01 5.42533863e-03
4.03703272e-01 -1.11405873e+00 -7.17809618e-01 1.40983880e-01
3.80416840e-01 -7.02090681e-01 -4.73092884e-01 1.58416837e-01
-6.00793958e-01 -1.00874078e+00 -9.38513160e-01 -3.24001700e-01
4.67458397e-01 -2.23662034e-01 7.71209002e-01 -2.68984102e-02
4.76563983e-02 9.81620848e-01 -6.15515411e-02 -1.02337563e+00
-6.96617961e-01 -4.40503538e-01 5.07241964e-01 6.11239851e-01
2.31322914e-01 -2.19791323e-01 -7.82265127e-01 4.55792308e-01
-1.13163924e+00 -5.86257696e-01 1.88832685e-01 4.47297305e-01
3.31428796e-01 5.92953086e-01 4.89446580e-01 -4.45684373e-01
1.23333669e+00 -4.34883535e-01 -1.10467482e+00 1.70336053e-01
-7.07987249e-01 1.85244635e-01 3.67994815e-01 -3.12945634e-01
-1.08572996e+00 -4.92744625e-01 6.82606399e-01 4.48749155e-01
2.94371158e-01 3.22469801e-01 4.17001724e-01 -1.80991873e-01
-3.61876875e-01 4.13322374e-02 3.88995439e-01 -5.60812950e-01
7.26346374e-02 6.40601933e-01 -1.07406988e-04 -3.40237170e-01
9.60985422e-01 9.37282443e-01 3.03239554e-01 -8.54887366e-01
-4.45273727e-01 -1.38377741e-01 -3.53724927e-01 -2.48154134e-01
6.88013017e-01 -6.04852796e-01 -1.68610811e+00 9.44990814e-01
-1.21828604e+00 -3.35592508e-01 -8.05112243e-01 6.16125405e-01
-8.43699813e-01 -8.63195285e-02 -1.09823608e+00 -1.58707786e+00
-4.57848124e-02 -4.87326205e-01 5.19646049e-01 5.08863568e-01
1.96503237e-01 -1.48323619e+00 6.47200584e-01 -7.20686167e-02
6.48938239e-01 -7.40001269e-04 7.94575155e-01 -8.23110759e-01
-6.16841495e-01 -2.63792664e-01 2.67609894e-01 4.70611155e-01
1.65091883e-02 1.02759816e-01 -2.95442402e-01 -1.17238954e-01
8.88644099e-01 4.79125917e-01 4.43852484e-01 7.84391999e-01
-9.59322229e-02 -3.67214620e-01 2.64144912e-02 -1.55368268e-01
1.48687148e+00 6.19048178e-01 -1.91897396e-02 7.31257617e-01
-3.02890301e-01 6.37787759e-01 4.22344029e-01 1.17185986e+00
4.80114996e-01 1.63417727e-01 4.88238662e-01 3.64513367e-01
8.51351976e-01 2.41406962e-01 5.59220076e-01 8.95131469e-01
-2.67514557e-01 -3.26579958e-01 -7.47416258e-01 3.53572875e-01
-1.91224897e+00 -1.14641929e+00 -2.19323620e-01 2.19608498e+00
3.71667713e-01 2.11791247e-02 4.78966773e-01 -2.83044606e-01
9.11361635e-01 -1.22698911e-01 -3.18380862e-01 -5.81491768e-01
-3.86354357e-01 1.08255208e-01 1.26910961e+00 9.23558235e-01
-5.66013217e-01 3.34745646e-01 7.13119888e+00 1.99250087e-01
-6.02309346e-01 5.50527237e-02 6.08103156e-01 -3.54755670e-01
-2.75104314e-01 1.13099165e-01 -7.88473964e-01 5.59790492e-01
9.35644209e-01 -8.83009374e-01 3.83592367e-01 2.24250644e-01
5.47196388e-01 -4.51723635e-01 -5.71010888e-01 4.84555572e-01
-2.10679397e-01 -9.83514249e-01 -3.48747224e-01 5.91897130e-01
6.66687250e-01 -1.18092529e-01 4.88983154e-01 -1.79844946e-01
8.31414223e-01 -2.52845138e-01 9.04135227e-01 1.03205681e+00
-2.62237877e-01 -8.55586588e-01 9.89381492e-01 4.73947078e-01
-9.99656022e-01 -3.57230842e-01 -2.93155015e-01 -6.01704895e-01
5.83739400e-01 3.77914935e-01 -1.81937125e-02 2.65871376e-01
6.46745801e-01 3.16007733e-01 -2.43165627e-01 9.35846388e-01
3.16315264e-01 4.69798297e-02 -3.90053540e-01 -5.63174129e-01
3.85196447e-01 -7.59297669e-01 6.84982240e-01 2.85587609e-01
3.45996678e-01 3.03946823e-01 -3.68618369e-01 7.00066864e-01
1.54842705e-01 5.33319041e-02 -2.15157971e-01 -1.21795155e-01
-2.83564001e-01 1.14528584e+00 -1.09388065e+00 -2.02527210e-01
-4.34961081e-01 8.92586172e-01 -5.24190962e-01 3.98519963e-01
-5.40454090e-01 -2.18158051e-01 7.50727713e-01 1.50694683e-01
6.69334471e-01 -6.79847598e-02 5.58355860e-02 -1.05679405e+00
1.43131893e-02 -4.45961028e-01 2.13127434e-01 -3.48540753e-01
-1.69860423e+00 5.67446351e-01 -5.22495173e-02 -8.10600042e-01
4.03944356e-03 -5.84208369e-01 -5.23203313e-01 7.12238133e-01
-1.29696190e+00 -4.14431602e-01 5.51970780e-01 4.73450869e-01
-2.41817422e-02 -5.57138503e-01 3.76731575e-01 -1.64728969e-01
-2.59620816e-01 -3.59924823e-01 9.05107617e-01 -7.76004121e-02
1.10419661e-01 -1.41423607e+00 3.44967902e-01 2.20073685e-01
-9.24384892e-02 4.56896931e-01 1.13929188e+00 -7.50930727e-01
-1.18235195e+00 -2.88780093e-01 9.01225448e-01 -6.04273379e-01
1.52565706e+00 -2.95237184e-01 -1.79919064e-01 5.89667916e-01
6.07616067e-01 -1.50171727e-01 7.22326756e-01 -4.26826924e-01
4.19277251e-01 -2.39228263e-01 -1.04408395e+00 4.70105946e-01
1.85771331e-01 -1.41118243e-02 -4.38013583e-01 1.76400900e-01
4.35047865e-01 3.97214115e-01 -6.76961601e-01 1.54285319e-02
7.23677933e-01 -1.14096904e+00 3.27880204e-01 -4.58636850e-01
-1.79929912e-01 3.53025692e-03 -6.57936558e-02 -1.46913946e+00
-9.40495357e-02 -1.07276273e+00 3.86870295e-01 9.51970696e-01
4.07952577e-01 -1.32970536e+00 6.43774688e-01 7.09827781e-01
8.96892130e-01 -3.91713142e-01 -1.04416430e+00 -1.10918760e+00
3.67755353e-01 2.94090778e-01 4.86258119e-01 4.83758390e-01
1.65422320e-01 1.91926122e-01 -1.30486503e-01 3.29161156e-03
1.04190850e+00 2.00518798e-02 1.37123331e-01 -1.38404000e+00
-4.10106152e-01 -4.96328503e-01 -2.83872217e-01 -7.28148997e-01
-3.13690938e-02 -3.01808149e-01 -2.07534492e-01 -9.51110721e-01
2.84756690e-01 -1.30634397e-01 -4.62509722e-01 -4.34036911e-01
5.76563060e-01 -1.55566916e-01 5.18622220e-01 2.81217992e-01
-6.59140289e-01 5.77729166e-01 1.16126049e+00 1.94332808e-01
-1.75910354e-01 6.94055915e-01 -7.06713259e-01 8.57600689e-01
7.23755002e-01 -7.35776305e-01 -4.33011688e-02 1.16391048e-01
8.12647521e-01 5.72831810e-01 2.63109803e-01 -4.11847889e-01
3.12309682e-01 -1.32202685e-01 -2.74793923e-01 -6.00915790e-01
7.16880113e-02 -9.75243092e-01 4.18119788e-01 8.59142900e-01
-4.44309741e-01 8.79374444e-01 -1.77641347e-01 8.94717038e-01
-2.42819771e-01 -5.71194470e-01 5.73589742e-01 -4.55921412e-01
1.10064775e-01 3.63297731e-01 -1.12731397e+00 -5.54103628e-02
1.40832722e+00 3.23777646e-01 -2.91859627e-01 -1.13082480e+00
-8.94870520e-01 4.35273260e-01 5.31111240e-01 2.05271646e-01
1.68499321e-01 -9.63124037e-01 -6.06454551e-01 -1.43944234e-01
-7.21852541e-01 -6.92431509e-01 1.10207386e-01 8.80811691e-01
-6.00040317e-01 2.52854884e-01 -2.18346834e-01 5.17130978e-02
-1.20733368e+00 7.66343594e-01 5.57512581e-01 -3.86393994e-01
-1.20689617e-02 5.07699132e-01 3.72426927e-01 5.28238639e-02
-1.36956275e-01 -2.08383664e-01 -1.84477836e-01 5.27673304e-01
7.34068811e-01 5.26420414e-01 -4.13967669e-01 -6.36090398e-01
-3.36813420e-01 6.08471215e-01 -2.56630510e-01 -6.37610734e-01
1.67988002e+00 -6.04996562e-01 -7.31948197e-01 1.15325570e+00
8.15237403e-01 1.57798544e-01 -1.02356195e+00 -1.78560968e-02
2.42400840e-01 5.55487201e-02 -3.31139326e-01 -4.45013583e-01
-9.64218199e-01 4.33236033e-01 2.97462285e-01 1.25206125e+00
5.90169787e-01 4.71518263e-02 5.84215701e-01 8.88230279e-02
3.43573034e-01 -1.24305356e+00 -1.28365770e-01 4.81483102e-01
6.49028182e-01 -7.44255126e-01 -2.25228950e-01 1.00725591e-01
-8.45622659e-01 9.12781656e-01 -2.05936298e-01 -5.33519447e-01
1.45211303e+00 7.04103708e-01 2.69761652e-01 -1.88386425e-01
-1.23810232e+00 -6.54956847e-02 -3.90409827e-01 4.08968359e-01
7.30585158e-02 2.05179721e-01 -6.17998004e-01 7.95983911e-01
-2.41608784e-01 -1.39233112e-01 1.30285823e+00 1.00893497e+00
-3.78796726e-01 -1.19981432e+00 -4.49450016e-01 1.10670008e-01
-1.06572902e+00 -5.53092994e-02 -3.97699982e-01 8.17587972e-01
-2.89426476e-01 8.19194674e-01 4.64513600e-01 3.44630480e-01
3.75741422e-01 -1.00865096e-01 3.59248310e-01 -2.92939901e-01
-6.66917443e-01 1.80624589e-01 -5.84067285e-01 -4.67476398e-02
-5.60097814e-01 -1.01910031e+00 -8.47531438e-01 -6.97447240e-01
-3.53939384e-01 7.35806406e-01 6.93923533e-01 8.48009348e-01
-2.37218469e-01 9.24542323e-02 8.54749382e-01 -3.05010974e-01
-1.17078698e+00 -5.52084804e-01 -1.79550648e+00 -5.71787283e-02
8.08960259e-01 -3.64412010e-01 -8.50167632e-01 -2.79710323e-01]
|
[4.951384544372559, 3.9483392238616943]
|
bb3e09ff-03a4-4113-9960-3b4f379fef7e
|
a-french-fairy-tale-corpus-syntactically-and
| null | null |
https://aclanthology.org/L12-1076
|
https://aclanthology.org/L12-1076.pdf
|
A French Fairy Tale Corpus syntactically and semantically annotated
|
Fairy tales, folktales and more generally children stories have lately attracted the Natural Language Processing (NLP) community. As such, very few corpora exist and linguistic resources are lacking. The work presented in this paper aims at filling this gap by presenting a syntactically and semantically annotated corpus. It focuses on the linguistic analysis of a Fairy Tales Corpus, and provides the description of the syntactic and semantic resources developed for Information Extraction. Resources include syntactic dependency relation annotation for 120 verbs; referential annotation, which is concerned with annotating each anaphoric occurrence and Proper Name with the most specific noun in the text; ontology matching for a substantial part of the nouns in the corpus; semantic role labelling for 41 verbs using the FrameNet database. The article also sums up previous analyses of this corpus and indicates possible uses of this corpus for the NLP community.
|
['Isma{\\"\\i}l El Maarouf', 'Jeanne Villaneau']
|
2012-05-01
| null | null | null |
lrec-2012-5
|
['ontology-matching']
|
['knowledge-base']
|
[ 7.17578158e-02 6.71203673e-01 -3.25476557e-01 -4.67993766e-01
-3.21738183e-01 -5.38078725e-01 6.62794709e-01 9.60550845e-01
-6.05204046e-01 1.00445080e+00 8.05934370e-01 6.43885210e-02
-5.37742853e-01 -7.76493430e-01 -1.50369346e-01 -3.00100058e-01
2.00225264e-01 8.89892995e-01 7.05755174e-01 -5.60937703e-01
4.80057150e-01 2.81870395e-01 -1.73221087e+00 5.15069723e-01
4.37375426e-01 5.57512283e-01 4.81317788e-01 5.02017103e-02
-7.27873027e-01 1.37469149e+00 -7.75364339e-01 -8.45511377e-01
-5.32440543e-01 -3.55608761e-01 -1.41650808e+00 -6.06467091e-02
8.60894397e-02 1.10072896e-01 9.04940665e-02 9.37662482e-01
3.57735157e-01 1.70306459e-01 2.99245358e-01 -1.04480481e+00
-3.10761571e-01 1.10651433e+00 3.64108644e-02 4.67429906e-01
7.85086215e-01 -6.97393179e-01 1.28235137e+00 -5.61352849e-01
1.35406578e+00 1.46916652e+00 3.40129942e-01 5.38646519e-01
-6.69184148e-01 -3.21977109e-01 -3.08170617e-01 4.43327546e-01
-1.06188059e+00 -3.37577850e-01 6.05853796e-01 -5.56213439e-01
1.57425654e+00 -3.31557612e-03 7.56270587e-01 7.81395853e-01
-1.44112080e-01 6.24286711e-01 8.33407640e-01 -1.13830912e+00
1.35403899e-02 1.15656070e-01 5.13933122e-01 4.61033076e-01
1.86303228e-01 -3.05287808e-01 -6.52397931e-01 -9.78306010e-02
5.73875070e-01 -7.39729524e-01 2.59265810e-01 -6.28923923e-02
-7.18059957e-01 9.26309764e-01 -2.51113236e-01 1.13241315e+00
-4.07690018e-01 3.00442763e-02 8.50971758e-01 2.25261271e-01
6.09792471e-01 4.46649343e-01 -5.40300429e-01 -5.68509161e-01
-3.20604831e-01 5.97957015e-01 1.07169676e+00 9.65141118e-01
2.41867930e-01 -3.50342304e-01 3.54796559e-01 1.19195831e+00
3.78549665e-01 8.94834846e-02 3.48347276e-01 -1.18007934e+00
5.16642988e-01 9.38255668e-01 2.46909186e-01 -9.24574494e-01
-5.65873086e-01 5.83081007e-01 4.20383632e-01 -2.94973016e-01
7.02061892e-01 9.57879946e-02 -1.25652894e-01 1.67914271e+00
5.78041852e-01 -7.02955484e-01 4.67562854e-01 4.63589400e-01
1.49351418e+00 3.87424171e-01 7.49747515e-01 -5.64470887e-01
1.93616307e+00 -5.63833237e-01 -1.36960757e+00 -4.20101769e-02
6.02022409e-01 -1.14475858e+00 7.59943247e-01 -5.16487025e-02
-1.41720653e+00 1.35892704e-01 -6.53803229e-01 -4.91446018e-01
-6.39045358e-01 -1.25918001e-01 8.09957504e-01 3.67064983e-01
-5.31737030e-01 2.28606135e-01 -7.10578561e-01 -1.02320910e+00
1.66580588e-01 2.74094790e-01 -4.31251049e-01 1.13394819e-02
-1.31716311e+00 1.44398844e+00 1.20364404e+00 -5.73227584e-01
-2.63108015e-02 -3.33763808e-01 -1.05014527e+00 -6.07414134e-02
7.89123714e-01 -1.14601023e-01 1.26977682e+00 -9.88596976e-01
-1.18888390e+00 1.66355515e+00 -4.67060842e-02 -4.62100387e-01
-1.03924565e-01 -1.69644490e-01 -5.85136950e-01 5.08756578e-01
7.36348927e-01 4.40362483e-01 -4.54682857e-02 -6.46993697e-01
-9.16509628e-01 -6.04150534e-01 4.18776333e-01 3.23508829e-01
-3.69225711e-01 1.38537645e+00 -3.82811457e-01 -9.06964540e-01
1.28332615e-01 -4.30606663e-01 2.75006324e-01 -4.65065390e-01
7.88018182e-02 -1.13178456e+00 2.72845536e-01 -6.46154404e-01
1.53922343e+00 -2.24098682e+00 -2.03201044e-02 -1.99349627e-01
3.72859277e-02 -4.49789800e-02 2.47570857e-01 1.02977550e+00
-2.28325665e-01 -2.23254204e-01 2.88114529e-02 2.25459754e-01
2.03532740e-01 8.21366549e-01 -2.22138584e-01 2.85806745e-01
-1.42740786e-01 6.60722852e-01 -1.00956893e+00 -9.62507248e-01
2.31839836e-01 3.07400018e-01 -2.70200133e-01 -2.65253991e-01
-3.17042768e-01 -1.86606497e-01 -6.64640665e-01 5.34940779e-01
-1.41538799e-01 3.07953358e-01 4.52314734e-01 8.00144896e-02
-7.23002493e-01 8.67289186e-01 -9.56451595e-01 1.38106334e+00
-1.16755418e-01 5.49925685e-01 -1.32608982e-02 -1.09865487e+00
8.46657693e-01 8.96890104e-01 6.14453435e-01 -6.39904022e-01
4.58660811e-01 4.74498659e-01 8.68464708e-02 -9.84777689e-01
5.20388007e-01 -2.63869315e-01 -4.33752179e-01 1.64541259e-01
2.38201633e-01 -2.24146366e-01 9.00049925e-01 1.31663069e-01
9.21708524e-01 5.21556854e-01 1.03644073e+00 -5.45385122e-01
6.84750974e-01 6.02968991e-01 6.58881724e-01 3.16286162e-02
-9.59463418e-02 -1.75518930e-01 6.25386178e-01 -8.24062169e-01
-9.05174315e-01 -6.80811644e-01 -6.46158159e-01 1.41518831e+00
1.12123275e-02 -7.67250478e-01 -9.04276788e-01 -2.60060519e-01
-5.35024226e-01 9.37641144e-01 -4.46416348e-01 5.36196470e-01
-8.36000979e-01 -6.00392580e-01 5.31393647e-01 3.82042289e-01
3.83123130e-01 -1.74867260e+00 -1.07959318e+00 6.32092118e-01
-4.59643543e-01 -1.32488263e+00 5.37830830e-01 3.31896506e-02
-4.53368396e-01 -1.51000011e+00 1.21858597e-01 -1.00641298e+00
1.22249238e-01 -5.67640960e-01 1.10740650e+00 1.06709458e-01
-1.18672654e-01 1.92113832e-01 -6.65807664e-01 -8.72076392e-01
-6.38903260e-01 -2.18319416e-01 -2.24998727e-01 -8.80228400e-01
1.18310189e+00 -6.14113510e-01 4.22839075e-01 -8.51257071e-02
-7.87515044e-01 -1.50370017e-01 -1.86125696e-01 4.53441262e-01
3.15108269e-01 1.78820118e-01 5.79959154e-01 -1.08757544e+00
4.45581645e-01 -4.23393756e-01 -5.29501021e-01 1.51659951e-01
-1.42608464e-01 -2.40824655e-01 6.47375286e-02 -5.96677624e-02
-1.40340984e+00 -2.59460688e-01 -5.23788214e-01 5.86472452e-01
-4.78325576e-01 3.53193581e-01 -4.11994576e-01 1.82298481e-01
6.73415184e-01 -3.13851595e-01 -3.98708135e-01 -5.98859727e-01
1.70492217e-01 4.81383592e-01 6.74124241e-01 -8.97821426e-01
1.20421886e-01 3.23527157e-01 1.39471754e-01 -1.07065427e+00
-1.12359893e+00 -6.73615575e-01 -7.20807016e-01 -2.89563775e-01
1.13118601e+00 -6.10705137e-01 -7.64887571e-01 -3.77666317e-02
-1.21662569e+00 7.22389296e-02 -5.98287463e-01 5.90590000e-01
-7.24505842e-01 2.32734904e-01 -7.19335496e-01 -7.09563911e-01
-1.91781610e-01 -6.19581580e-01 6.74351931e-01 1.25867695e-01
-8.04266512e-01 -1.08356667e+00 1.67015772e-02 6.82365000e-01
-1.10492818e-01 5.26302457e-01 1.30148113e+00 -1.04923368e+00
2.73014635e-01 -3.65799367e-02 -6.83590770e-03 -1.66857108e-01
-2.37653345e-01 -1.96167842e-01 -5.33979952e-01 3.51848990e-01
7.61969909e-02 -5.36088347e-01 1.42172500e-01 1.14040785e-01
1.78232729e-01 -4.13737774e-01 -2.97938943e-01 -3.17294151e-01
1.64832199e+00 5.63549936e-01 5.78421175e-01 8.88995886e-01
2.82452703e-01 1.35604584e+00 9.13768589e-01 2.31557071e-01
5.87684214e-01 7.69151688e-01 5.88907972e-02 7.06660211e-01
-1.95811808e-01 -5.55194058e-02 2.08443739e-02 5.60140967e-01
-4.50389683e-01 -9.25094634e-02 -1.02422440e+00 9.54358459e-01
-1.91893458e+00 -1.00489855e+00 -4.20173556e-01 1.49513829e+00
1.03883243e+00 -1.70930009e-02 1.14245802e-01 4.49107826e-01
7.87705421e-01 4.82931733e-04 3.65437806e-01 -4.24622625e-01
-2.20777124e-01 5.98261178e-01 -9.35978144e-02 5.58717787e-01
-1.04923761e+00 1.49091959e+00 5.91594505e+00 9.35683489e-01
-1.43267080e-01 5.10366797e-01 -3.38222325e-01 2.73447782e-01
4.44389731e-02 2.07705066e-01 -9.04018223e-01 3.43605518e-01
8.45213711e-01 -2.32725099e-01 -6.16851076e-02 7.22776711e-01
2.30678782e-01 -4.73938376e-01 -8.57362986e-01 5.28446019e-01
1.60757378e-01 -1.46350825e+00 -2.73007989e-01 -1.95879474e-01
2.71914840e-01 -1.47038281e-01 -8.91707361e-01 1.05994299e-01
2.64525831e-01 -5.80407917e-01 1.10692394e+00 1.22473337e-01
4.82752353e-01 -7.36190856e-01 8.50185752e-01 2.48974234e-01
-1.05090511e+00 5.98966377e-04 -3.77907544e-01 -4.66983050e-01
2.79169530e-01 3.30422148e-02 -3.37879747e-01 2.45252699e-01
8.44366610e-01 4.87156957e-01 -3.43105912e-01 6.57247484e-01
-9.04209673e-01 5.92422068e-01 -5.04283309e-01 -3.12392175e-01
2.36402646e-01 -1.27744213e-01 8.82369578e-01 1.17871940e+00
-2.02455327e-01 6.81195617e-01 1.95436135e-01 4.64345932e-01
1.79895744e-01 8.38956773e-01 -3.33141774e-01 -1.43820181e-01
8.26168239e-01 8.85067165e-01 -1.22515988e+00 -3.02615881e-01
-6.75346911e-01 4.22262698e-01 2.34599724e-01 -2.24939525e-01
-2.44162470e-01 -5.30124605e-01 2.80187547e-01 2.68164605e-01
5.34622893e-02 2.98758775e-01 -5.41636767e-03 -7.77597964e-01
-1.19249865e-01 -5.05488217e-01 8.58161926e-01 -8.74191105e-01
-1.09819651e+00 3.96995634e-01 7.24575639e-01 -3.59399199e-01
-4.95052785e-01 -4.69080448e-01 -1.66944027e-01 4.50876504e-01
-9.72964108e-01 -1.36987126e+00 2.98243672e-01 4.22907799e-01
5.94327033e-01 -1.37159407e-01 1.25554585e+00 3.27675432e-01
-1.94100812e-01 -4.23255526e-02 -4.20752168e-01 2.73117989e-01
4.09810722e-01 -1.06722105e+00 -1.76327035e-01 4.33853656e-01
-2.21316591e-01 5.36892951e-01 1.07410240e+00 -7.97249019e-01
-5.14181912e-01 -2.83898145e-01 2.12082076e+00 -4.14812684e-01
1.21210480e+00 1.29955590e-01 -8.10413718e-01 8.61639202e-01
5.00439644e-01 -4.38868135e-01 8.14444780e-01 -4.77275848e-02
-4.27058302e-02 5.32571554e-01 -1.44488311e+00 2.75524586e-01
1.03605819e+00 -2.72893190e-01 -1.65221512e+00 5.11339724e-01
6.62481546e-01 -5.28134704e-01 -1.12575614e+00 8.67361948e-02
3.26745957e-01 -6.68847859e-01 8.50387573e-01 -5.11541963e-01
5.30632019e-01 -9.16281492e-02 -3.06421578e-01 -4.71476048e-01
9.57311466e-02 -3.22849780e-01 4.36309636e-01 1.77624476e+00
1.45715460e-01 -2.75117636e-01 4.51739699e-01 6.03355825e-01
-3.40293616e-01 -2.73130745e-01 -1.26602685e+00 -2.88169712e-01
3.84497307e-02 -6.09604776e-01 4.09327686e-01 1.17915010e+00
7.16674626e-01 7.43922591e-01 2.43511111e-01 -2.86085159e-01
5.55161059e-01 1.83950644e-02 7.52223358e-02 -1.65862083e+00
2.99578253e-02 -3.97706538e-01 -5.62061727e-01 -4.71157394e-02
4.96466815e-01 -8.72709453e-01 -1.62439048e-01 -1.70322311e+00
8.06896761e-02 -1.36828870e-01 6.29193306e-01 8.93446326e-01
3.15140218e-01 1.92524776e-01 8.88503268e-02 2.32848212e-01
-5.43791234e-01 1.42950654e-01 8.17576647e-01 4.61530954e-01
-1.08645111e-01 -3.15339476e-01 -4.98055816e-01 1.43725979e+00
6.77002192e-01 -7.54974544e-01 -1.23298474e-01 -1.72420606e-01
7.22168207e-01 3.54622602e-02 3.86765450e-02 -5.44231415e-01
3.43259200e-02 -3.40702623e-01 -1.35512561e-01 -4.99866992e-01
1.26356497e-01 -8.55029047e-01 1.00074641e-01 2.50786185e-01
-2.50907809e-01 -9.42832530e-02 2.82901019e-01 -9.87697914e-02
-3.83318722e-01 -1.16348374e+00 7.45981812e-01 -4.18198943e-01
-1.19425189e+00 -4.89821076e-01 -8.24154437e-01 4.16170806e-01
1.17891574e+00 -8.29583630e-02 -2.12051958e-01 1.77872166e-01
-1.06277680e+00 1.23041324e-01 3.46915632e-01 3.37976426e-01
2.42355883e-01 -1.13981247e+00 -4.74051982e-01 -5.10457039e-01
2.63296247e-01 -1.84763491e-01 3.54893319e-02 3.80049884e-01
-7.80303776e-01 4.42969769e-01 -3.86687726e-01 2.57250875e-01
-1.66705501e+00 5.64175904e-01 -1.62865147e-02 -1.51879206e-01
-8.23756039e-01 6.45338058e-01 -2.56162137e-01 -1.91081062e-01
1.40154690e-01 1.16209723e-01 -1.24674606e+00 4.59424883e-01
4.91293997e-01 6.17210329e-01 -2.99137563e-01 -1.42022240e+00
-4.44905639e-01 3.71700764e-01 3.18465531e-01 -3.37850064e-01
1.81797552e+00 -3.65535289e-01 -9.95149076e-01 5.76751351e-01
6.89275384e-01 1.62456959e-01 -2.74666041e-01 -2.61955738e-01
7.04571009e-01 -2.44296029e-01 -2.35284418e-01 -6.12206876e-01
-4.63734418e-01 2.90036708e-01 -1.02570867e-02 4.97234851e-01
7.90809810e-01 7.32210875e-01 4.44754601e-01 3.49066377e-01
3.11399043e-01 -1.50957251e+00 -2.34381288e-01 7.79558539e-01
8.77640486e-01 -5.58541238e-01 1.58058479e-01 -1.11588299e+00
-3.81226242e-01 1.29528844e+00 3.89369965e-01 2.80597452e-02
4.08346266e-01 4.89861101e-01 -9.66068581e-02 -1.06229603e+00
-4.55455869e-01 -5.26242197e-01 -1.73095450e-01 5.61439037e-01
8.27543437e-01 9.05536860e-03 -1.67892182e+00 9.18306947e-01
-6.92410529e-01 2.16894597e-02 6.38430953e-01 1.13271224e+00
-7.32527316e-01 -1.38954091e+00 -4.30635363e-01 -1.15954161e-01
-1.28706455e+00 -1.62281170e-01 -5.62600911e-01 1.29515648e+00
6.02149844e-01 1.06826890e+00 2.57145524e-01 6.60058796e-01
4.52205688e-01 2.57789016e-01 8.55347574e-01 -1.05036247e+00
-8.15277338e-01 2.59465665e-01 9.64311540e-01 -4.81460214e-01
-1.31775844e+00 -8.54858041e-01 -1.75084889e+00 -1.39974073e-01
-1.89602315e-01 5.22529185e-01 5.07880747e-01 1.44653177e+00
-6.71073020e-01 3.17473590e-01 -4.06445146e-01 -3.56069595e-01
1.97648227e-01 -9.31328714e-01 -6.74820662e-01 3.56700420e-01
-5.74863076e-01 -9.28687155e-01 4.89424802e-02 2.55435824e-01]
|
[10.063883781433105, 9.40151596069336]
|
4fcb7ecd-a34b-40fe-9713-e5df1d727f67
|
coverage-of-information-extraction-from
| null | null |
https://aclanthology.org/D19-1583
|
https://aclanthology.org/D19-1583.pdf
|
Coverage of Information Extraction from Sentences and Paragraphs
|
Scalar implicatures are language features that imply the negation of stronger statements, e.g., {``}She was married twice{''} typically implicates that she was not married thrice. In this paper we discuss the importance of scalar implicatures in the context of textual information extraction. We investigate how textual features can be used to predict whether a given text segment mentions all objects standing in a certain relationship with a certain subject. Preliminary results on Wikipedia indicate that this prediction is feasible, and yields informative assessments.
|
['Paramita Mirza', 'Nitisha Jain', 'Simon Razniewski', 'Gerhard Weikum']
|
2019-11-01
| null | null | null |
ijcnlp-2019-11
|
['implicatures']
|
['natural-language-processing']
|
[ 1.48330435e-01 8.59365642e-01 -5.85184455e-01 -9.03491020e-01
-4.19372082e-01 -7.34656572e-01 9.02118742e-01 9.28400636e-01
-4.28214222e-01 1.33290470e+00 6.49055541e-01 -5.79074144e-01
-2.66101241e-01 -8.35932851e-01 -7.40851820e-01 -1.50975659e-01
-1.89518794e-01 4.32743132e-01 1.10256799e-01 -4.92339134e-01
2.91416675e-01 3.91644627e-01 -1.41668200e+00 7.58041382e-01
5.76170266e-01 6.32376373e-01 6.40548440e-03 1.00893728e-01
-3.38988572e-01 1.30384755e+00 -9.00872290e-01 -1.12832761e+00
-2.72630185e-01 -2.67168949e-03 -1.03981102e+00 -9.05590970e-03
7.62109160e-01 -3.21606189e-01 -8.76352638e-02 1.12127793e+00
-2.53964007e-01 -2.14640439e-01 9.90369380e-01 -1.31440437e+00
-5.80445051e-01 1.40862036e+00 -6.49091899e-01 6.94200814e-01
8.22217286e-01 -2.14620814e-01 1.86006618e+00 -7.24188864e-01
5.88532031e-01 1.27872801e+00 6.35263503e-01 -2.59350780e-02
-1.11993504e+00 -5.47046125e-01 2.81577528e-01 8.71456340e-02
-1.24191499e+00 -6.36785448e-01 6.87569559e-01 -3.08557719e-01
1.40166664e+00 7.17450917e-01 3.80835742e-01 8.75067711e-01
4.20852214e-01 9.57393289e-01 8.33340764e-01 -5.50660610e-01
-1.95858702e-01 3.83542299e-01 4.86124009e-01 6.85922861e-01
1.18082166e+00 -4.24579531e-01 -6.34117424e-01 -4.28250164e-01
1.52789399e-01 -5.74399590e-01 6.48847446e-02 3.08467448e-01
-1.38385105e+00 9.32276368e-01 3.30575034e-02 7.17661321e-01
-3.74396294e-01 2.30935529e-01 3.42197865e-01 2.49637052e-01
4.91434515e-01 8.70675445e-01 -8.73603642e-01 -9.35046896e-02
-7.70491779e-01 5.55994511e-01 1.11606979e+00 9.90965724e-01
6.70508564e-01 -3.76484573e-01 1.40607700e-01 4.37908828e-01
3.78619492e-01 3.35108817e-01 6.38191104e-02 -1.26350665e+00
7.73731947e-01 6.82368219e-01 4.52820092e-01 -1.18581545e+00
-5.19963503e-01 -1.90624148e-01 -3.25427204e-01 -2.45348796e-01
5.32256603e-01 -3.14204454e-01 2.27282364e-02 1.54188871e+00
8.68899524e-02 -1.88089624e-01 4.27170932e-01 5.19874394e-01
1.04095602e+00 5.69419503e-01 4.98495221e-01 -4.82375354e-01
1.59268630e+00 1.25541519e-02 -1.26031744e+00 -6.16520107e-01
9.47581470e-01 -7.63867915e-01 9.30076301e-01 3.08871835e-01
-1.10705400e+00 2.74273846e-02 -1.01983619e+00 -2.32570440e-01
-2.75017828e-01 3.85314077e-01 1.37225890e+00 5.91312289e-01
-3.57358009e-01 6.73106730e-01 -6.15534186e-01 -2.73959935e-01
4.61746156e-02 3.59520644e-01 -7.14106023e-01 3.26541245e-01
-1.55716956e+00 1.17852330e+00 6.69096470e-01 1.93389311e-01
2.25227475e-02 -1.88647404e-01 -1.23435771e+00 1.20201066e-01
4.52869058e-01 -4.12224501e-01 1.27428949e+00 -9.35561597e-01
-6.62618697e-01 1.26806426e+00 -3.07371318e-01 -6.09180570e-01
2.27789506e-01 -3.48946035e-01 -8.68475199e-01 2.44262666e-02
5.33740580e-01 1.84705704e-01 3.98196667e-01 -8.56777430e-01
-8.77963901e-01 -6.34351015e-01 7.15289414e-01 3.47830147e-01
-5.22529244e-01 6.09655321e-01 7.06597567e-02 -4.68958110e-01
3.75241935e-01 -6.66057289e-01 7.72317648e-02 -2.34079763e-01
-1.02782369e+00 -7.74043500e-01 6.14413261e-01 -2.39188999e-01
1.46701884e+00 -1.92216372e+00 -3.10925871e-01 1.44305050e-01
1.26624197e-01 -1.98804155e-01 5.37040532e-01 3.74407649e-01
-3.11590731e-01 5.77251613e-01 1.74511820e-01 3.10320199e-01
3.41276765e-01 6.11465573e-01 -4.11835581e-01 7.62124300e-01
4.21951443e-01 6.06130898e-01 -7.77314544e-01 -8.41036916e-01
-1.70845792e-01 -6.02560565e-02 -4.23823178e-01 -1.44117549e-01
-4.01498616e-01 -5.12779117e-01 -6.37245536e-01 4.68948215e-01
3.55382979e-01 -2.92665124e-01 7.08942294e-01 -5.07973731e-01
-3.94969761e-01 9.03683484e-01 -1.19124484e+00 8.71625602e-01
5.79963215e-02 7.34503925e-01 3.31946164e-02 -8.68649542e-01
7.01759160e-01 3.52385581e-01 1.38219625e-01 -1.84193030e-01
1.71917543e-01 3.77189010e-01 1.68255299e-01 -8.08769763e-01
7.59995818e-01 -4.97377545e-01 -5.15052021e-01 3.10128897e-01
-3.49465162e-01 -4.17893440e-01 6.99563503e-01 5.73040307e-01
9.63287413e-01 -1.38508365e-01 1.01033640e+00 -6.01999938e-01
6.62095189e-01 8.45180899e-02 7.38632739e-01 5.18417418e-01
6.17970899e-02 3.74961793e-02 1.09202373e+00 -2.94475943e-01
-8.86332929e-01 -1.04114330e+00 -3.34675878e-01 8.94883037e-01
1.50190428e-01 -6.18807018e-01 -3.94741036e-02 -7.07688868e-01
2.66270190e-01 1.52526879e+00 -8.83406103e-01 2.37380013e-01
-7.18243241e-01 -5.81379592e-01 5.76831579e-01 6.58173978e-01
-3.34032848e-02 -5.94018042e-01 -5.79125941e-01 -8.13311059e-03
-4.92090166e-01 -1.18250012e+00 -2.42702905e-02 4.59981382e-01
-6.43339694e-01 -1.02844524e+00 1.17796034e-01 -6.15106821e-01
6.46939158e-01 -3.95986766e-01 1.34083211e+00 5.96778274e-01
1.58946693e-01 -1.18064210e-01 -2.22348467e-01 -4.47678298e-01
-3.47186863e-01 -1.27666384e-01 2.84448743e-01 -4.03202862e-01
7.20804632e-01 -2.86244869e-01 7.77311847e-02 -1.88139036e-01
-7.32372224e-01 -2.94265389e-01 3.46434444e-01 5.47841907e-01
1.82451576e-01 3.45971555e-01 1.68730676e-01 -1.58716953e+00
6.86338961e-01 -6.54262900e-01 -1.33891270e-01 4.25757259e-01
-2.71567822e-01 2.65298903e-01 3.45767736e-01 -1.58899263e-01
-1.34278989e+00 -1.72245607e-01 -3.38327259e-01 6.19259000e-01
-2.40365446e-01 1.10540223e+00 -4.45557505e-01 6.23819411e-01
9.07301128e-01 -2.22788557e-01 -3.91318291e-01 -4.51922417e-03
1.20046049e-01 5.42312145e-01 4.03368294e-01 -9.12773311e-01
6.41744256e-01 4.10862684e-01 1.45427465e-01 -8.83830607e-01
-1.34982955e+00 -2.35447928e-01 -6.76461577e-01 -4.55218414e-03
5.36356211e-01 -8.58004987e-01 -8.93286467e-01 -4.01079804e-01
-1.35656106e+00 2.34626219e-01 -8.82355347e-02 5.95533192e-01
-4.12516147e-01 4.38862026e-01 -6.07161403e-01 -8.85486364e-01
1.48704872e-01 -6.03858471e-01 7.63146341e-01 -9.85142589e-02
-1.07804263e+00 -1.00075817e+00 -3.75472367e-01 2.97000915e-01
-3.24026376e-01 3.11578095e-01 1.21691573e+00 -1.20429265e+00
-6.04262054e-02 -2.46762037e-01 -4.55980003e-02 -7.77009130e-02
3.41763347e-01 5.40178895e-01 -4.55217183e-01 1.69810593e-01
-7.89015219e-02 -3.65419239e-01 2.78028011e-01 1.21757738e-01
7.05043256e-01 -1.07622659e+00 -5.23671269e-01 -7.74342045e-02
1.28864408e+00 -2.13442385e-01 4.96278942e-01 4.81366992e-01
5.69146909e-02 9.58616734e-01 9.81342375e-01 5.14742017e-01
4.89022344e-01 4.77183193e-01 1.78663358e-01 4.35659736e-01
6.19809963e-02 -8.39431211e-02 3.17944229e-01 1.08433478e-01
4.33687940e-02 -4.13735509e-01 -9.16940510e-01 6.36375368e-01
-1.44955468e+00 -1.35317373e+00 -9.42623496e-01 1.48907471e+00
1.29569137e+00 6.10987782e-01 -3.14029604e-01 4.38775748e-01
5.58277965e-01 1.25539154e-01 2.08249819e-02 -4.73557681e-01
-4.27771360e-01 -3.80372941e-01 1.70005947e-01 7.97402740e-01
-9.60237265e-01 8.93411994e-01 6.72083521e+00 3.81408155e-01
-6.09503090e-01 -9.89358723e-02 5.14451385e-01 1.70058787e-01
-9.99010146e-01 1.46534786e-01 -1.16278732e+00 1.65004209e-01
7.13366151e-01 -6.46970689e-01 -4.75699335e-01 7.17123330e-01
-5.03445826e-02 -4.28660363e-01 -1.45105672e+00 2.52602488e-01
1.26844436e-01 -1.25672650e+00 1.37568220e-01 -1.54715568e-01
2.75041968e-01 -4.02666271e-01 1.31882280e-01 1.15072116e-01
3.66299957e-01 -7.66804457e-01 1.07514262e+00 1.59243122e-01
5.56235731e-01 -6.97462499e-01 8.53555381e-01 5.38307250e-01
-6.89879000e-01 -5.25414757e-02 -2.83792883e-01 -4.52879310e-01
2.75281042e-01 9.43603933e-01 -1.14324617e+00 2.42090553e-01
5.00154376e-01 3.78864527e-01 -5.07072031e-01 4.81273890e-01
-8.39203477e-01 5.92436373e-01 -5.06772280e-01 -5.44222951e-01
2.12909445e-01 1.01780713e-01 6.63865387e-01 1.45820141e+00
5.44548556e-02 6.50250137e-01 -1.18878134e-01 8.20092022e-01
9.42147076e-02 1.82345435e-01 -8.19442809e-01 -2.06925571e-01
3.87314528e-01 9.36207473e-01 -7.13449419e-01 -5.02276242e-01
-6.74966693e-01 5.09196401e-01 5.16931593e-01 2.06700265e-01
-6.01324022e-01 -4.03739244e-01 5.14433205e-01 3.71812552e-01
6.40893504e-02 -1.82822496e-02 -5.29728174e-01 -1.16618288e+00
-8.46842583e-03 -5.00363231e-01 7.28670418e-01 -9.16046262e-01
-1.35426486e+00 4.38348919e-01 1.86837375e-01 -6.95451915e-01
-5.56843579e-01 -8.88003945e-01 -4.94203031e-01 5.09107172e-01
-1.25200868e+00 -9.80158746e-01 3.67247909e-01 3.47747713e-01
1.84198171e-01 9.08242539e-02 9.21787381e-01 -2.17415005e-01
-1.89590454e-01 2.72433966e-01 -6.83873236e-01 2.35335335e-01
4.11461502e-01 -1.46779883e+00 1.12729527e-01 9.23348188e-01
3.00201058e-01 1.18337989e+00 1.72841144e+00 -9.57093775e-01
-1.26398969e+00 -5.81447303e-01 2.06048346e+00 -7.48029411e-01
1.12792146e+00 3.58386077e-02 -7.57322848e-01 1.34660017e+00
3.60670418e-01 -1.40958307e-02 1.01697290e+00 7.29002416e-01
-4.12654817e-01 -4.78852764e-02 -1.20751107e+00 5.89121222e-01
7.44232655e-01 -5.27250588e-01 -1.56924951e+00 6.32425606e-01
5.06676316e-01 -2.36488268e-01 -7.81110466e-01 5.59434950e-01
4.01828319e-01 -8.47814798e-01 8.24647188e-01 -9.52599764e-01
4.79707509e-01 -2.21895546e-01 -3.89004827e-01 -8.35745454e-01
-3.49704355e-01 -2.31558070e-01 -3.01978230e-01 1.18179178e+00
1.15323925e+00 -3.71653855e-01 7.37936556e-01 1.20922410e+00
3.95241641e-02 -3.68491232e-01 -1.05520022e+00 -7.29706347e-01
2.28883147e-01 -5.30973136e-01 5.07921576e-01 1.30916798e+00
1.10650885e+00 7.11238086e-01 4.13272902e-02 2.81802446e-01
5.10245562e-01 3.05723310e-01 1.68103367e-01 -1.24027121e+00
-1.76284701e-01 -2.17147693e-01 -4.37203705e-01 -7.96571195e-01
5.64131439e-01 -8.90783012e-01 -8.82310495e-02 -1.61946726e+00
3.02329332e-01 -3.34937781e-01 1.83665410e-01 6.51703596e-01
-1.85437590e-01 2.00185906e-02 -3.31272274e-01 -1.89867467e-01
-5.17349243e-01 1.37737364e-01 7.82051980e-01 -3.23964983e-01
4.26310658e-01 1.06149778e-01 -1.14512312e+00 1.16566670e+00
7.00889647e-01 -3.52628499e-01 -2.09693998e-01 -4.37678427e-01
1.19587886e+00 3.48308384e-01 4.19090278e-02 -3.03484708e-01
3.29933643e-01 -5.03763437e-01 3.93495977e-01 -4.94313478e-01
3.60106319e-01 -1.03925192e+00 -2.78139673e-02 3.98686111e-01
-6.12298906e-01 2.35952169e-01 2.11593062e-01 2.97737956e-01
-3.77548426e-01 -7.77479589e-01 6.20077290e-02 -1.54155076e-01
-6.25811219e-01 -2.35001892e-01 -4.98590261e-01 1.06205672e-01
8.65487099e-01 1.62127450e-01 -3.58922333e-01 -3.29688758e-01
-7.17057705e-01 -7.13944137e-02 1.02737956e-01 1.79198280e-01
7.04345703e-01 -1.07020807e+00 -8.53514791e-01 -4.31746960e-01
2.80874640e-01 -5.33185065e-01 -4.42737669e-01 6.16430223e-01
-3.56637597e-01 4.42842394e-01 1.00187853e-01 -2.51188837e-02
-1.57299006e+00 3.75526160e-01 -1.15404308e-01 2.52269190e-02
-4.07327861e-01 1.05768096e+00 -4.41988707e-02 -2.30614841e-01
-1.68927182e-02 -4.58239555e-01 -6.59630299e-01 1.76162675e-01
6.04938686e-01 -1.26266824e-02 -7.99727738e-02 -8.51164162e-01
-7.21695662e-01 -2.38761202e-01 -2.58739233e-01 -1.94184706e-01
1.48697317e+00 -7.65504390e-02 -8.40377808e-01 8.97690117e-01
1.02285552e+00 6.71276510e-01 -5.56119263e-01 -2.48976126e-02
3.44947010e-01 -4.86879021e-01 -2.57644057e-01 -9.51287329e-01
-5.36010742e-01 2.39912942e-01 -6.83630288e-01 6.86662555e-01
4.96123731e-01 6.52543962e-01 1.40686274e-01 7.92857170e-01
3.40670139e-01 -9.66056526e-01 -3.39489341e-01 5.46954036e-01
1.15106213e+00 -1.00659931e+00 7.62676477e-01 -1.10636234e+00
-6.12696886e-01 1.32413852e+00 7.17206955e-01 1.54454291e-01
5.43527424e-01 6.41591251e-01 -4.21462953e-01 -5.90465844e-01
-1.03708243e+00 1.50979355e-01 2.24359110e-01 6.92477599e-02
1.20494545e+00 3.61629516e-01 -9.98874843e-01 7.73809612e-01
-7.14371622e-01 -3.54908466e-01 8.49481523e-01 6.94220185e-01
-5.61423898e-01 -8.37147892e-01 -1.91495776e-01 7.56535232e-01
-8.75444174e-01 -4.30175781e-01 -5.28863966e-01 8.49011540e-01
1.97073311e-01 1.12159741e+00 -5.81624880e-02 2.76324935e-02
1.73747972e-01 -1.32466584e-01 7.24585056e-01 -8.70710194e-01
-4.95944798e-01 -3.82338673e-01 1.02027464e+00 -1.95141107e-01
-6.63624525e-01 -1.37811899e+00 -1.65632343e+00 -3.75821233e-01
-5.06464183e-01 5.47431946e-01 2.62970597e-01 1.33088243e+00
-3.30769718e-01 -5.60113713e-02 -4.55161445e-02 1.89058617e-01
-5.50222993e-01 -1.05992877e+00 -6.73581839e-01 3.68628144e-01
3.71020824e-01 -5.29698133e-01 -6.76859200e-01 9.37145129e-02]
|
[9.980610847473145, 8.92947006225586]
|
19b79a59-5f87-4702-aff6-ae2a1b3613b6
|
cross-domain-text-classification-with
| null | null |
https://aclanthology.org/P16-1155
|
https://aclanthology.org/P16-1155.pdf
|
Cross-domain Text Classification with Multiple Domains and Disparate Label Sets
| null |
['Shourya Roy', 'Himanshu Sharad Bhatt', 'Manjira Sinha']
|
2016-08-01
| null | null | null |
acl-2016-8
|
['cross-domain-text-classification']
|
['natural-language-processing']
|
[-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01]
|
[-7.463138580322266, 3.6788830757141113]
|
6ff54e4d-4b33-4ab6-a66d-e28ddce5c439
|
solving-mixed-modal-jigsaw-puzzle-for-fine
| null | null |
http://openaccess.thecvf.com/content_CVPR_2020/html/Pang_Solving_Mixed-Modal_Jigsaw_Puzzle_for_Fine-Grained_Sketch-Based_Image_Retrieval_CVPR_2020_paper.html
|
http://openaccess.thecvf.com/content_CVPR_2020/papers/Pang_Solving_Mixed-Modal_Jigsaw_Puzzle_for_Fine-Grained_Sketch-Based_Image_Retrieval_CVPR_2020_paper.pdf
|
Solving Mixed-Modal Jigsaw Puzzle for Fine-Grained Sketch-Based Image Retrieval
|
ImageNet pre-training has long been considered crucial by the fine-grained sketch-based image retrieval (FG-SBIR) community due to the lack of large sketch-photo paired datasets for FG-SBIR training. In this paper, we propose a self-supervised alternative for representation pre-training. Specifically, we consider the jigsaw puzzle game of recomposing images from shuffled parts. We identify two key facets of jigsaw task design that are required for effective FG-SBIR pre-training. The first is formulating the puzzle in a mixed-modality fashion. Second we show that framing the optimisation as permutation matrix inference via Sinkhorn iterations is more effective than the common classifier formulation of Jigsaw self-supervision. Experiments show that this self-supervised pre-training strategy significantly outperforms the standard ImageNet-based pipeline across all four product-level FG-SBIR benchmarks. Interestingly it also leads to improved cross-category generalisation across both pre-train/fine-tune and fine-tune/testing stages.
|
[' Yi-Zhe Song', ' Tao Xiang', ' Timothy M. Hospedales', ' Yongxin Yang', 'Kaiyue Pang']
|
2020-06-01
| null | null | null |
cvpr-2020-6
|
['sketch-based-image-retrieval']
|
['computer-vision']
|
[ 5.27857482e-01 -3.07001434e-02 -3.19075793e-01 -2.94189513e-01
-1.04700124e+00 -8.93631399e-01 9.91893768e-01 -1.95312560e-01
-2.73045838e-01 3.96601528e-01 3.70083153e-01 -2.08203509e-01
-5.52926064e-01 -6.75094604e-01 -9.64285970e-01 -3.93234670e-01
3.08433235e-01 8.03043246e-01 2.93284655e-02 -3.25673282e-01
4.58220869e-01 4.36252803e-01 -1.60863483e+00 8.16530883e-01
2.20390916e-01 1.11559498e+00 2.69190788e-01 7.31665969e-01
-2.11997598e-01 8.30116093e-01 -3.70923102e-01 -7.71470189e-01
5.89499712e-01 -2.67770588e-01 -1.11109245e+00 2.63090014e-01
9.43198323e-01 -2.19827592e-01 -1.26336053e-01 6.25612617e-01
3.55386496e-01 1.07943393e-01 5.74432731e-01 -1.35982001e+00
-7.99650252e-01 5.71906924e-01 -6.21294320e-01 -1.05512895e-01
2.86408037e-01 3.17809314e-01 1.65689814e+00 -9.22954679e-01
1.04357898e+00 1.36447525e+00 6.59451365e-01 3.51714224e-01
-1.53998029e+00 -5.58757484e-01 -3.17920223e-02 3.15137088e-01
-1.22614372e+00 -3.29548568e-01 8.09681237e-01 -4.56436127e-01
9.20012355e-01 2.88716376e-01 4.73163933e-01 1.20841587e+00
-4.61234480e-01 1.04751086e+00 1.28277659e+00 -5.31034291e-01
6.54281378e-02 -9.93323922e-02 -1.27260700e-01 6.03987336e-01
-3.80104482e-02 1.62116453e-01 -7.02961087e-01 1.68777108e-02
1.14208281e+00 -1.66537270e-01 2.35703304e-01 -7.17832088e-01
-1.41642451e+00 8.39249730e-01 6.13310218e-01 3.14158708e-01
-2.23920226e-01 4.70629901e-01 6.34511173e-01 6.77403867e-01
2.82174557e-01 9.56541896e-01 -4.14489597e-01 6.81356043e-02
-1.13258207e+00 3.96395952e-01 5.85339725e-01 8.74056041e-01
9.33442652e-01 -3.10873002e-01 -3.78649652e-01 1.12839413e+00
-2.96610668e-02 5.56844696e-02 5.02874469e-03 -1.07831502e+00
4.90585506e-01 4.08314019e-01 -1.62807658e-01 -8.44711602e-01
5.95563799e-02 -3.45178157e-01 -6.97429776e-01 2.83623308e-01
6.07061088e-01 6.71832025e-01 -1.00491667e+00 1.44350970e+00
1.74223911e-02 -6.63677305e-02 -1.67288616e-01 9.59508181e-01
9.15359735e-01 3.74170750e-01 2.67213404e-01 5.22751272e-01
1.45526135e+00 -1.39987516e+00 -1.71025053e-01 -3.41806680e-01
2.15893522e-01 -9.72129703e-01 1.43973517e+00 6.10860109e-01
-9.92250204e-01 -6.99986398e-01 -1.02587056e+00 -3.25436085e-01
-5.10274231e-01 4.25538957e-01 8.49879026e-01 4.50159669e-01
-1.05328846e+00 8.68057728e-01 -7.97699243e-02 -3.94325733e-01
8.62207353e-01 1.82494506e-01 -8.93062234e-01 -6.56104267e-01
-6.86469018e-01 7.34540403e-01 4.48212922e-01 -6.55058548e-02
-8.69654417e-01 -9.98267114e-01 -4.93157536e-01 -1.22144431e-01
6.00970685e-01 -8.77432942e-01 1.06167376e+00 -1.21050942e+00
-1.37369251e+00 1.41164720e+00 3.42757076e-01 -1.85843125e-01
6.32288516e-01 -3.21254544e-02 1.94280386e-01 1.85165673e-01
1.65787563e-01 1.27442837e+00 1.23456287e+00 -1.34456062e+00
-3.85197878e-01 -2.73117036e-01 3.45623881e-01 1.67351007e-01
1.67478040e-01 -7.23450556e-02 -7.72058189e-01 -9.30702448e-01
-9.80465114e-02 -9.38775063e-01 -8.01085681e-02 1.55899554e-01
-2.70261914e-01 -3.62745941e-01 5.06387711e-01 -4.80956882e-01
5.52609622e-01 -1.96900785e+00 2.19736770e-01 2.91094780e-01
1.06128126e-01 2.93463171e-01 -7.32118011e-01 6.51902735e-01
-3.56775314e-01 6.04400272e-03 -7.43543613e-04 -5.46811402e-01
2.41504729e-01 3.54429156e-01 -4.67200130e-01 1.32530376e-01
5.89738548e-01 1.44821072e+00 -9.54140902e-01 -3.46080929e-01
3.48477781e-01 2.76425451e-01 -5.45310140e-01 2.04575837e-01
-5.70782721e-01 3.23044032e-01 -1.18239783e-01 7.74041235e-01
6.18567407e-01 -4.14560050e-01 1.52092308e-01 -6.11096740e-01
1.31014943e-01 1.55870214e-01 -1.05141675e+00 2.25870132e+00
-6.14263177e-01 6.52976155e-01 -5.76687232e-02 -1.09673750e+00
9.55832899e-01 -1.74405992e-01 3.18527579e-01 -1.01864207e+00
-3.54102887e-02 1.92696735e-01 -4.55834836e-01 -1.30751267e-01
5.43002725e-01 -2.94299603e-01 -3.29157449e-02 6.46432161e-01
5.52695572e-01 -3.56570035e-01 2.97461689e-01 2.25931197e-01
1.20472550e+00 6.52105510e-01 1.23545773e-01 -2.60285705e-01
4.80170131e-01 2.14157358e-01 -1.70160253e-02 9.37480211e-01
3.28279197e-01 1.04607499e+00 4.40799981e-01 -5.52690446e-01
-1.34061515e+00 -1.11569870e+00 2.69136578e-01 1.47165489e+00
7.61436522e-02 -5.18607974e-01 -4.87798423e-01 -7.15876639e-01
1.90359309e-01 3.19567263e-01 -9.05101895e-01 1.52160615e-01
-4.77388203e-01 -2.58471698e-01 6.11938894e-01 3.76680613e-01
5.43459475e-01 -1.40572667e+00 -3.33822042e-01 4.96368222e-02
-4.43306984e-03 -1.12199676e+00 -5.02297044e-01 8.49670246e-02
-4.77449298e-01 -1.28193748e+00 -8.44171524e-01 -7.55596757e-01
5.33229828e-01 4.27247196e-01 1.57026577e+00 3.23519439e-01
-7.46106923e-01 6.64183974e-01 -5.34156859e-01 1.95452701e-02
-2.92075515e-01 2.28228867e-01 -6.47264361e-01 -1.34380367e-02
1.01625562e-01 -6.02957547e-01 -6.84556365e-01 4.43992585e-01
-9.63528633e-01 2.99415708e-01 1.10190344e+00 9.00926590e-01
6.56960487e-01 -2.70691901e-01 3.99367630e-01 -9.24242198e-01
6.29263222e-01 -9.17286947e-02 -5.49043834e-01 5.88782370e-01
-4.80349481e-01 2.81263560e-01 3.34794939e-01 -4.29205030e-01
-7.28737772e-01 1.09343819e-01 -4.37154137e-02 -7.42793381e-01
-2.17818484e-01 2.48811781e-01 -3.52424532e-02 -3.44319224e-01
6.59092069e-01 -5.19136675e-02 1.56080753e-01 -6.00681067e-01
7.81967998e-01 2.71519631e-01 6.52105331e-01 -9.45365727e-01
9.62308884e-01 4.65922803e-01 1.42171025e-01 -7.34966457e-01
-8.80137265e-01 -8.98908854e-01 -4.83627170e-01 -1.77328244e-01
8.29181194e-01 -8.89210939e-01 -7.95121908e-01 2.19756991e-01
-1.11341178e+00 -5.83635271e-01 -5.67472577e-01 -1.60522938e-01
-7.77636766e-01 2.94359803e-01 -3.27544749e-01 -5.44210911e-01
-3.82562131e-01 -1.19509554e+00 1.47288287e+00 -1.94055304e-01
-1.17687784e-01 -7.43654728e-01 1.97475359e-01 7.39651024e-01
4.53148574e-01 1.48885950e-01 7.56472588e-01 -3.90744060e-01
-8.05330932e-01 6.53454289e-02 -8.72336924e-01 3.37329835e-01
-1.55706286e-01 -2.96893358e-01 -1.10092294e+00 -1.53176889e-01
-6.52130425e-01 -7.16208160e-01 1.12649381e+00 -5.10620289e-02
1.19319630e+00 -2.05825597e-01 1.30545124e-01 6.38000607e-01
1.73346925e+00 -5.92058480e-01 9.44085836e-01 5.14145374e-01
8.24361801e-01 7.89295316e-01 8.36138666e-01 7.63102770e-02
3.50591153e-01 9.39085841e-01 4.33665037e-01 -1.35235578e-01
-6.27289474e-01 -5.14427662e-01 8.73263553e-02 2.38115057e-01
-2.40461566e-02 4.33310196e-02 -6.76385701e-01 6.80420160e-01
-1.90892434e+00 -1.00961280e+00 -1.80252008e-02 2.09956789e+00
7.41357267e-01 -1.61122620e-01 4.80061263e-01 1.14110045e-01
3.01295429e-01 2.24166662e-01 -1.87385648e-01 -2.32703134e-01
-2.59864599e-01 6.58088267e-01 7.41083860e-01 3.64241123e-01
-1.30197072e+00 1.17207718e+00 5.60908890e+00 1.28563023e+00
-8.16975355e-01 2.36199588e-01 7.13926494e-01 1.53483823e-01
-3.54357302e-01 7.81947896e-02 -4.70109105e-01 3.45282331e-02
2.39020869e-01 5.06317019e-01 8.63959491e-01 8.63714516e-01
-4.97187227e-01 -2.28184074e-01 -1.29183805e+00 1.51183021e+00
2.55171806e-01 -1.75787091e+00 9.02840868e-02 -1.30129950e-02
8.13928604e-01 5.98996803e-02 -2.65380163e-02 2.40948632e-01
3.81789237e-01 -1.19468570e+00 1.00560987e+00 2.90206134e-01
1.09864140e+00 -3.99283797e-01 3.88733476e-01 -2.82139108e-02
-1.13303041e+00 -1.02693886e-01 -3.54790568e-01 1.72957152e-01
-8.87841135e-02 3.81924242e-01 -8.13220084e-01 7.27482200e-01
6.33946538e-01 6.29836619e-01 -9.57315564e-01 8.83943200e-01
-3.16444367e-01 2.56465077e-01 -2.00942785e-01 2.65154660e-01
4.94302064e-01 -2.00212404e-01 2.74898291e-01 1.25104797e+00
-4.03197072e-02 -3.54481220e-01 8.04779530e-02 9.82014596e-01
1.99444909e-02 -6.82824478e-02 -5.08956969e-01 -2.08678663e-01
-4.46346849e-02 1.58670378e+00 -8.53545785e-01 -2.02264972e-02
-1.20588161e-01 1.23147166e+00 4.45260108e-01 1.03400268e-01
-4.92628932e-01 -1.96492244e-02 5.05285203e-01 1.81402579e-01
7.95154870e-01 -1.39670283e-01 -3.38968843e-01 -1.03562593e+00
-6.41832873e-02 -9.31833565e-01 4.45242077e-01 -1.09637046e+00
-1.70828521e+00 3.24988097e-01 5.20467162e-02 -9.84636605e-01
-1.81283280e-01 -8.67843866e-01 -5.26488900e-01 6.21276140e-01
-1.68002844e+00 -1.99423933e+00 -4.17671144e-01 5.77361524e-01
6.01236761e-01 -1.40883639e-01 7.95790136e-01 3.60664129e-01
-8.32767934e-02 7.20394969e-01 -3.54701936e-01 3.11122485e-03
7.46235013e-01 -1.35745716e+00 6.65284693e-01 5.24355471e-01
6.63957536e-01 4.57728893e-01 5.57456017e-01 -5.32784998e-01
-1.73509026e+00 -1.05830181e+00 4.79710042e-01 -5.22081733e-01
7.00820863e-01 -6.65218174e-01 -6.21673465e-01 3.59371632e-01
2.25061908e-01 1.24893546e-01 2.21283853e-01 2.58000523e-01
-1.03916025e+00 -2.11296007e-01 -9.24106479e-01 3.98639619e-01
1.36059856e+00 -7.79052258e-01 -4.10355330e-01 4.75936741e-01
3.31620663e-01 -8.85144770e-02 -8.38715017e-01 2.00595036e-01
8.33550751e-01 -8.57103348e-01 1.50646257e+00 -7.35260129e-01
8.96902502e-01 -1.92544103e-01 -3.33903223e-01 -1.14737391e+00
-4.39658016e-01 -7.39046276e-01 4.37912554e-01 1.25768054e+00
2.09223554e-01 -7.40556791e-02 1.09645784e+00 2.02365398e-01
1.07894287e-01 -4.19275552e-01 -6.26678824e-01 -8.78644347e-01
-1.45383418e-01 -4.86595392e-01 4.81369942e-01 8.32797468e-01
-3.74535471e-01 5.45130670e-01 -4.62494969e-01 -3.74185920e-01
7.67203391e-01 1.98213458e-01 1.31200194e+00 -1.01899016e+00
-6.58246040e-01 -7.24697948e-01 -5.19628406e-01 -9.11407053e-01
2.78121829e-02 -1.06102014e+00 9.91586521e-02 -1.57247794e+00
2.50001997e-01 -8.46482515e-01 -2.21706375e-01 7.05448031e-01
-2.29698028e-02 7.97139227e-01 4.52719331e-01 1.98234946e-01
-9.33161497e-01 1.17101006e-01 1.22398150e+00 -2.65782446e-01
2.23720312e-01 -3.83547187e-01 -6.98679328e-01 3.58207747e-02
5.95020354e-01 -2.52617836e-01 -5.51132202e-01 -3.58530819e-01
5.65552175e-01 -2.40214929e-01 8.72973204e-01 -5.85820556e-01
4.16452251e-02 -9.45466980e-02 1.63656995e-01 -5.50255477e-01
3.72762144e-01 -7.30231285e-01 4.40463066e-01 -1.03011280e-01
-4.93551701e-01 -5.18785529e-02 2.24742055e-01 5.81981957e-01
-1.08803079e-01 -2.87105799e-01 5.56934834e-01 -5.00647366e-01
-9.71490681e-01 1.76016361e-01 2.34467417e-01 1.02830432e-01
6.05940938e-01 -3.61203402e-01 -3.72772813e-01 -3.31270307e-01
-5.03457010e-01 -1.07445657e-01 5.09628534e-01 6.63227379e-01
4.61907804e-01 -1.36997163e+00 -6.27895474e-01 2.44533807e-01
5.07028580e-01 -1.46429762e-01 4.24258739e-01 7.09079087e-01
-4.96220827e-01 4.81956303e-01 -5.12539744e-01 -6.57990515e-01
-1.30578673e+00 5.58653057e-01 -7.45717883e-02 -7.46342182e-01
-5.36889017e-01 9.65627134e-01 2.86189675e-01 -6.97952926e-01
1.86516255e-01 1.21037923e-01 1.97514579e-01 1.17592067e-01
3.97095084e-01 1.64431721e-01 3.80699337e-01 -5.65816283e-01
-2.01151416e-01 5.77604115e-01 -9.76293311e-02 -3.15608859e-01
1.68249393e+00 2.75123775e-01 -1.88084364e-01 1.10139176e-01
1.33963108e+00 -3.49356651e-01 -1.18551910e+00 -2.07905650e-01
1.97936535e-01 -6.94214165e-01 9.50911790e-02 -1.18633318e+00
-1.00567031e+00 7.67994940e-01 2.54648387e-01 -2.29692496e-02
8.86899412e-01 2.62772620e-01 3.37324411e-01 4.75469887e-01
4.58784699e-01 -1.02965176e+00 5.59651494e-01 2.80990034e-01
1.35580492e+00 -1.25104105e+00 2.67037153e-01 -3.98986667e-01
-6.68215334e-01 1.12721241e+00 3.26646209e-01 -5.68520546e-01
3.09694231e-01 -2.23350912e-01 -1.83958307e-01 -5.38950741e-01
-5.15696228e-01 -4.71336246e-01 7.17288017e-01 5.31550705e-01
6.75347224e-02 -2.61121914e-02 1.30207419e-01 3.87209535e-01
-3.38701420e-02 1.10543445e-01 -1.06215421e-02 7.51243711e-01
1.33393958e-01 -1.49768329e+00 -3.02966774e-01 5.37095010e-01
-9.29053575e-02 -7.77599141e-02 -7.38000154e-01 9.67628598e-01
2.11629737e-02 5.38487613e-01 -1.16242711e-02 -3.07537168e-01
2.74965525e-01 -2.68828385e-02 1.17207396e+00 -5.86826980e-01
-9.26716864e-01 -2.50225924e-02 3.14635634e-01 -8.35015357e-01
-5.82680821e-01 -4.55842257e-01 -3.96078467e-01 -1.21158510e-01
-2.70217210e-01 -2.54580677e-01 9.31407750e-01 9.39057946e-01
5.04212856e-01 2.15985090e-01 3.82678837e-01 -1.26198208e+00
-3.73669803e-01 -8.23933423e-01 -3.44303936e-01 7.65917003e-01
-4.39734533e-02 -8.70148301e-01 -1.90874413e-01 -5.82588725e-02]
|
[11.614635467529297, 0.5508106350898743]
|
da21b9a7-14c5-452f-8446-d3acad8687be
|
pre-and-post-counting-for-scalable
|
2110.09767
| null |
https://arxiv.org/abs/2110.09767v1
|
https://arxiv.org/pdf/2110.09767v1.pdf
|
Pre and Post Counting for Scalable Statistical-Relational Model Discovery
|
Statistical-Relational Model Discovery aims to find statistically relevant patterns in relational data. For example, a relational dependency pattern may stipulate that a user's gender is associated with the gender of their friends. As with propositional (non-relational) graphical models, the major scalability bottleneck for model discovery is computing instantiation counts: the number of times a relational pattern is instantiated in a database. Previous work on propositional learning utilized pre-counting or post-counting to solve this task. This paper takes a detailed look at the memory and speed trade-offs between pre-counting and post-counting strategies for relational learning. A pre-counting approach computes and caches instantiation counts for a large set of relational patterns before model search. A post-counting approach computes an instantiation count dynamically on-demand for each candidate pattern generated during the model search. We describe a novel hybrid approach, tailored to relational data, that achieves a sweet spot with pre-counting for patterns involving positive relationships (e.g. pairs of users who are friends) and post-counting for patterns involving negative relationships (e.g. pairs of users who are not friends). Our hybrid approach scales model discovery to millions of data facts.
|
['Oliver Schulte', 'Richard Mar']
|
2021-10-19
| null | null | null | null |
['model-discovery']
|
['miscellaneous']
|
[-3.30951577e-03 3.92567039e-01 -5.68146169e-01 -5.13198197e-01
-5.68389535e-01 -2.74947196e-01 7.57239103e-01 1.02990139e+00
-1.83403343e-01 7.20791280e-01 -2.71680772e-01 -6.31733119e-01
-4.73411441e-01 -1.58677137e+00 -7.60023594e-01 -1.15074933e-01
-7.75369763e-01 1.43630147e+00 6.22533858e-01 2.72482574e-01
1.47778578e-02 6.55848980e-01 -1.91193652e+00 7.07683027e-01
5.33233136e-02 8.32529187e-01 -3.08006704e-01 6.06213629e-01
-4.12328571e-01 1.13507104e+00 -3.80219489e-01 -7.12453246e-01
-1.06406920e-02 -1.77712798e-01 -1.01152849e+00 -4.14988995e-01
4.06356812e-01 1.02050871e-01 -2.48036921e-01 8.39223444e-01
-3.84485833e-02 -5.39966226e-02 5.42116642e-01 -1.64039075e+00
2.44270325e-01 1.06093860e+00 -5.79093874e-01 4.90550488e-01
7.08824575e-01 -5.13824463e-01 1.42594814e+00 -7.33286440e-01
6.12389326e-01 1.54609513e+00 5.67378163e-01 1.66727111e-01
-1.48878658e+00 -7.29875863e-01 1.02053978e-01 3.92344296e-01
-1.60830534e+00 -3.22701633e-01 3.40748250e-01 -3.05272073e-01
1.20682800e+00 8.95937324e-01 7.02164590e-01 4.72089529e-01
-1.72658995e-01 5.49113750e-01 8.81851673e-01 -6.23816788e-01
3.60416085e-01 3.34002018e-01 4.46713358e-01 9.19856489e-01
5.05983949e-01 4.40423191e-02 -9.34733748e-01 -8.27327669e-01
6.62393153e-01 -5.84835820e-02 5.71840823e-01 -2.56464392e-01
-7.21658528e-01 7.93195724e-01 -1.37490770e-02 3.42599362e-01
-1.10632069e-01 3.20723981e-01 3.17788750e-01 3.80589813e-01
1.62586242e-01 4.47035342e-01 -8.56034994e-01 -1.93318740e-01
-9.83838320e-01 4.53790724e-01 1.36876571e+00 1.15033662e+00
1.15310180e+00 -7.29421437e-01 -6.65658861e-02 6.53595030e-01
4.83644903e-01 9.43124574e-03 -4.95188534e-02 -6.45852983e-01
3.49733204e-01 8.92204344e-01 -2.28940230e-02 -1.17348826e+00
-5.15780866e-01 -1.77654579e-01 -5.84131062e-01 -5.80520868e-01
4.88010556e-01 2.44342804e-01 -2.57323235e-01 1.63504553e+00
7.21568346e-01 2.52249122e-01 -2.51944005e-01 4.45288718e-02
8.20918858e-01 5.43575764e-01 1.64637893e-01 -6.88702583e-01
1.51796961e+00 -3.88499767e-01 -3.82638544e-01 -5.67813739e-02
8.98264349e-01 -5.51953256e-01 6.67556763e-01 3.19404811e-01
-1.10252142e+00 -2.11964697e-01 -5.34650683e-01 2.45965421e-02
-5.47033429e-01 -1.69361532e-01 1.19750762e+00 5.20387650e-01
-6.19315267e-01 5.30704081e-01 -8.75082731e-01 -3.39467824e-01
2.33143985e-01 6.90139651e-01 -3.12398583e-01 3.83532569e-02
-1.15692067e+00 9.58094180e-01 5.01063466e-01 -2.19503179e-01
-2.68540174e-01 -8.47592711e-01 -7.13753998e-01 1.11159556e-01
8.11231375e-01 -6.13605320e-01 1.27839935e+00 -4.23415244e-01
-7.43490815e-01 1.02531588e+00 -6.85733020e-01 -5.92108786e-01
4.91136312e-01 1.13067940e-01 -5.25241077e-01 -2.53867716e-01
1.71875879e-01 1.57068774e-01 2.19097972e-01 -1.35008156e+00
-9.23214495e-01 -4.10925984e-01 2.50385672e-01 -1.63994730e-01
-1.48541471e-02 2.34975621e-01 -8.60098779e-01 9.17613972e-03
4.77079660e-01 -6.84918404e-01 -2.46286213e-01 -4.09894019e-01
-5.80523491e-01 -8.67487848e-01 6.70010507e-01 8.63018334e-02
1.72699130e+00 -1.64972639e+00 -1.45583838e-01 1.09685159e+00
3.88595253e-01 -1.62744999e-01 3.84476781e-01 7.22068667e-01
-2.91732460e-01 3.22221130e-01 1.72274113e-01 -3.16097349e-01
-4.18439433e-02 6.04996204e-01 -4.16518599e-01 2.77324587e-01
3.59859504e-02 6.13983810e-01 -8.51921022e-01 -1.06613183e+00
1.44107100e-02 -4.88484539e-02 -6.08041644e-01 -2.35432629e-02
-6.14326656e-01 -4.70470011e-01 -2.14484692e-01 7.39598274e-01
4.35800910e-01 -4.05634552e-01 7.02204823e-01 -1.48386717e-01
9.59950220e-03 7.15972424e-01 -1.57405508e+00 5.73854089e-01
-4.24605936e-01 1.19834520e-01 -2.56390870e-01 -6.66032016e-01
8.49158227e-01 -3.84831391e-02 4.62767541e-01 -6.54426277e-01
-1.41301200e-01 1.44895494e-01 -1.29064843e-01 -3.20237786e-01
4.61272806e-01 -3.18065733e-01 -2.04861596e-01 6.68988168e-01
-1.89342678e-01 1.66970049e-03 6.58430815e-01 4.35875475e-01
1.39789534e+00 -2.75074482e-01 6.37255013e-01 -1.58453628e-01
5.41297436e-01 9.42676812e-02 4.69968110e-01 9.55594122e-01
7.08627641e-01 -3.75650940e-03 1.24537468e+00 -7.24156260e-01
-6.66272879e-01 -9.29886341e-01 -1.17872864e-01 1.17343724e+00
-9.60214287e-02 -1.32910001e+00 -9.80775580e-02 -7.41197348e-01
2.97545552e-01 5.28592050e-01 -7.41141975e-01 5.78869246e-02
-6.27975583e-01 -5.99764943e-01 2.81127274e-01 3.80776852e-01
1.65544562e-02 -7.80997336e-01 -4.30700898e-01 2.16876730e-01
8.87925923e-02 -9.44006324e-01 3.72707218e-01 4.86894578e-01
-8.17293763e-01 -1.44243526e+00 7.57251322e-01 -4.88277733e-01
8.44865978e-01 -1.52786851e-01 1.54323936e+00 5.10974884e-01
-2.21456885e-01 1.32398903e-01 -1.52324531e-02 -4.48813170e-01
-3.32874060e-01 7.92523846e-02 -1.02282697e-02 -2.34865457e-01
8.67561340e-01 -6.33179069e-01 1.65997863e-01 2.96645075e-01
-6.23804152e-01 -5.01497537e-02 4.15109307e-01 4.08793926e-01
1.01381338e+00 7.11669922e-01 -1.44051453e-02 -1.51398337e+00
4.78429079e-01 -8.29993784e-01 -8.38823795e-01 5.08278906e-01
-6.00872099e-01 4.68724146e-02 3.50913614e-01 -4.85185981e-01
-3.83858383e-01 2.37615764e-01 2.33244777e-01 -2.56518930e-01
1.08398579e-01 8.97545576e-01 -8.21068063e-02 6.57196641e-01
5.74437022e-01 -1.16202042e-01 -3.31599981e-01 -3.34791958e-01
8.21414962e-02 2.70843953e-01 2.59590745e-01 -1.10126138e+00
8.68901014e-01 1.99195281e-01 6.08623266e-01 -8.58351111e-01
-8.21032822e-01 -7.32152998e-01 -7.11882889e-01 -5.43271611e-03
1.83198914e-01 -7.15438366e-01 -1.25330245e+00 -1.00172162e-01
-9.64777172e-01 -3.09940994e-01 -4.04215157e-01 1.29170790e-01
-3.05833519e-01 9.48059745e-03 -2.07982123e-01 -1.02369201e+00
1.16666213e-01 -3.58732134e-01 7.78086722e-01 -1.52255595e-01
-8.94348383e-01 -9.16066587e-01 1.69111431e-01 1.11279272e-01
-9.45585743e-02 1.71973690e-01 1.50712597e+00 -1.02145433e+00
-7.22821951e-01 -5.84579885e-01 -2.81613797e-01 -5.28043985e-01
-1.10110261e-01 1.02308132e-01 -4.78820235e-01 6.98859021e-02
-5.66453695e-01 -1.05906077e-01 3.48507285e-01 -1.62684768e-01
1.28054857e+00 -7.23449588e-01 -8.60282063e-01 -1.07817613e-01
1.23580694e+00 5.19192917e-03 4.57191467e-01 2.11114679e-02
6.45334303e-01 6.75455391e-01 6.30049109e-01 5.28372467e-01
7.05855489e-01 7.84377158e-01 8.23271871e-02 -2.81535331e-02
3.77823025e-01 -6.42240763e-01 -2.06194207e-01 6.58159703e-02
-1.56872392e-01 1.78371415e-01 -1.24836767e+00 4.21988189e-01
-1.97180474e+00 -1.19749224e+00 -5.96105576e-01 2.43910003e+00
1.38019145e+00 5.96854448e-01 4.33144927e-01 3.99852008e-01
3.50078613e-01 -3.03701043e-01 -2.20607221e-01 -4.74859327e-01
3.02508116e-01 5.21260142e-01 3.15394133e-01 9.60779607e-01
-9.26285148e-01 9.59585428e-01 6.18300247e+00 8.22833657e-01
-5.22667646e-01 2.77521815e-02 5.43630064e-01 -3.81942689e-01
-4.46576834e-01 3.55427504e-01 -1.36199641e+00 1.80546165e-01
1.13716972e+00 -2.58752853e-01 2.28946939e-01 1.10824680e+00
-1.93904936e-01 -6.03855073e-01 -1.83623338e+00 9.43420529e-01
-1.51384160e-01 -1.43075180e+00 9.76681560e-02 3.63558859e-01
4.12309438e-01 -2.14118108e-01 -3.96639764e-01 5.84579289e-01
5.78043401e-01 -1.01231897e+00 7.07032919e-01 6.17544234e-01
4.96087432e-01 -9.90983903e-01 5.33036709e-01 5.53032219e-01
-1.32578802e+00 -1.25112742e-01 -2.27784723e-01 -2.92003542e-01
-2.49327600e-01 1.00474644e+00 -1.12853277e+00 2.61036515e-01
7.49266982e-01 4.20325190e-01 -6.64235115e-01 6.95265174e-01
-4.86904643e-02 4.63205129e-01 -9.53671157e-01 -1.55760661e-01
-1.81081027e-01 2.46839412e-02 5.19083887e-02 1.43103266e+00
-6.26953319e-02 8.29216018e-02 1.51362181e-01 7.18499959e-01
-1.04238406e-01 7.33053684e-03 -6.27899945e-01 1.49599999e-01
9.30515528e-01 1.03755343e+00 -7.86941290e-01 -5.38361430e-01
-4.69373882e-01 3.98045219e-03 5.65535367e-01 -4.47763242e-02
-6.82317853e-01 -3.02255061e-03 3.33726019e-01 9.12816823e-01
2.07347423e-01 -1.75301790e-01 -2.18233079e-01 -7.45416939e-01
-1.34582072e-01 -8.36929262e-01 8.74752760e-01 -3.50125700e-01
-1.23164189e+00 1.03714108e-01 6.64710760e-01 -6.67263091e-01
-4.39692169e-01 -2.61769652e-01 -4.36599612e-01 6.70832872e-01
-8.94740522e-01 -1.05208814e+00 1.26325712e-01 5.81443012e-01
-7.65396059e-02 2.07561981e-02 1.06580949e+00 2.26763979e-01
-4.21279013e-01 6.04639113e-01 -6.99328363e-01 -7.94918835e-02
2.58719511e-02 -1.34246659e+00 9.45679620e-02 5.47833920e-01
5.01765072e-01 1.19005501e+00 7.32379436e-01 -7.54947305e-01
-1.40338707e+00 -9.89264488e-01 1.77629566e+00 -6.57648683e-01
7.68669963e-01 -5.27923465e-01 -8.80338550e-01 1.03328252e+00
-4.83379871e-01 1.85083911e-01 1.07753158e+00 1.18878543e+00
-6.83715522e-01 -4.13026214e-01 -1.11816823e+00 6.14555180e-01
1.00523710e+00 -5.76323092e-01 -4.42035198e-01 4.33461726e-01
1.94510549e-01 -3.82687271e-01 -9.50972915e-01 3.03635210e-01
5.21462977e-01 -8.40488911e-01 9.15795147e-01 -7.48946428e-01
3.42795819e-01 -3.73711050e-01 -2.70667970e-01 -3.68100703e-01
-1.53725654e-01 -8.55401516e-01 -8.80885959e-01 1.17976034e+00
8.34694207e-01 -3.80038083e-01 9.39065695e-01 9.00289476e-01
5.11945069e-01 -1.14021957e+00 -1.07420778e+00 -6.13620996e-01
-3.46019775e-01 -6.41760826e-01 7.35412896e-01 1.01279640e+00
2.79540449e-01 5.52910805e-01 -1.42615989e-01 3.35920662e-01
5.17795205e-01 4.20500815e-01 1.02299392e+00 -1.65067625e+00
-5.30646324e-01 -3.48069489e-01 -3.84259462e-01 -5.33057988e-01
-4.72602434e-02 -9.36122835e-01 -5.50064862e-01 -1.54015160e+00
4.52677369e-01 -8.80492032e-01 1.01894297e-01 9.56457615e-01
2.45592728e-01 -4.24863659e-02 -2.70054966e-01 2.43579939e-01
-7.93352604e-01 -3.12636316e-01 6.57853663e-01 1.94020554e-01
-4.00102794e-01 3.88840556e-01 -6.32603407e-01 8.33311439e-01
6.10829413e-01 -9.03360665e-01 -4.19031441e-01 2.30922580e-01
1.30464637e+00 3.09670120e-01 3.68136197e-01 -7.39520133e-01
6.75326884e-01 -3.25670362e-01 3.18190634e-01 -1.14885259e+00
3.57241154e-01 -8.45542967e-01 5.21679103e-01 4.32736605e-01
-3.64949763e-01 2.07993254e-01 2.04876155e-01 5.57063103e-01
-1.79890841e-01 -2.59180844e-01 4.84367728e-01 -3.33468527e-01
-1.03166580e-01 1.36009306e-01 -2.93082237e-01 3.60888280e-02
1.22571492e+00 -1.47490367e-01 -9.11669880e-02 3.46357897e-02
-1.09809053e+00 3.13804448e-01 3.46876420e-02 1.69197589e-01
4.52052861e-01 -1.18066680e+00 -2.78704375e-01 2.04253927e-01
1.95850357e-01 3.78288656e-01 -3.52880657e-01 6.53832674e-01
-1.95631459e-01 4.02002364e-01 4.36558932e-01 -4.07835662e-01
-1.94718587e+00 4.73031968e-01 3.29486698e-01 -6.93291962e-01
-1.62751034e-01 1.03923106e+00 -5.63486218e-01 -5.29120207e-01
3.68969887e-01 -4.89109695e-01 -3.02938893e-02 5.03768981e-01
4.69691008e-01 4.98868883e-01 2.27437645e-01 -3.71538728e-01
-5.79475701e-01 2.88839996e-01 -3.61023337e-01 4.67395484e-02
1.47542417e+00 4.27943408e-01 -8.49039376e-01 8.74000847e-01
9.21987414e-01 1.89788625e-01 -4.28942621e-01 -5.73793173e-01
4.64761227e-01 -4.77422655e-01 -1.95983335e-01 -7.26003468e-01
-7.67438352e-01 2.64047980e-01 -1.37875646e-01 7.00107813e-01
5.88129103e-01 8.40819120e-01 9.65133980e-02 7.08355069e-01
7.27184713e-01 -9.16561306e-01 4.39350288e-06 5.66570044e-01
6.94429457e-01 -8.36980879e-01 5.76475382e-01 -7.71856725e-01
-1.42099425e-01 9.00203764e-01 7.12082863e-01 2.00640619e-01
9.96559739e-01 5.26933789e-01 -6.70641303e-01 -6.16090894e-01
-1.28775024e+00 -6.89172819e-02 4.85167354e-02 1.52572095e-01
3.88801694e-01 3.83164674e-01 -2.02444911e-01 6.03754997e-01
-6.87078953e-01 -3.12207807e-02 2.00959280e-01 1.05834663e+00
-2.65177071e-01 -1.54064906e+00 -2.32026696e-01 9.71869469e-01
-3.67489457e-01 -1.25564098e-01 -5.69732666e-01 1.00114644e+00
5.04144549e-01 7.55001485e-01 3.76682341e-01 -3.71462464e-01
4.80387926e-01 1.06567271e-01 9.27594781e-01 -1.17026472e+00
-6.17588997e-01 -1.63248926e-01 5.64848483e-01 -5.44828236e-01
-2.77667820e-01 -9.71746445e-01 -1.44017541e+00 -6.42754197e-01
-2.71082520e-01 3.33515644e-01 2.79497325e-01 9.12806094e-01
2.25710422e-02 5.41949943e-02 3.05690974e-01 1.53820338e-02
-3.03239495e-01 -7.45157063e-01 -5.97642899e-01 -1.09474873e-02
3.01484782e-02 -6.88795626e-01 -2.46011525e-01 -9.24695358e-02]
|
[9.024761199951172, 7.270232200622559]
|
aa5d89b6-0b19-4175-8c3d-36db12379797
|
unsupervised-domain-adaptation-on-question
| null | null |
https://aclanthology.org/2022.sigdial-1.42
|
https://aclanthology.org/2022.sigdial-1.42.pdf
|
Unsupervised Domain Adaptation on Question-Answering System with Conversation Data
|
Machine reading comprehension (MRC) is a task for question answering that finds answers to questions from documents of knowledge. Most studies on the domain adaptation of MRC require documents describing knowledge of the target domain. However, it is sometimes difficult to prepare such documents. The goal of this study was to transfer an MRC model to another domain without documents in an unsupervised manner. Therefore, unlike previous studies, we propose a domain-adaptation framework of MRC under the assumption that the only available data in the target domain are human conversations between a user asking questions and an expert answering the questions. The framework consists of three processes: (1) training an MRC model on the source domain, (2) converting conversations into documents using document generation (DG), a task we developed for retrieving important information from several human conversations and converting it to an abstractive document text, and (3) transferring the MRC model to the target domain with unsupervised domain adaptation. To the best of our knowledge, our research is the first to use conversation data to train MRC models in an unsupervised manner. We show that the MRC model successfully obtains question-answering ability from conversations in the target domain.
|
['Yasuhiro Sogawa', 'Takeshi Homma', 'Amalia Adiba']
| null | null | null | null |
sigdial-acl-2022-9
|
['machine-reading-comprehension']
|
['natural-language-processing']
|
[ 7.35299110e-01 6.00688815e-01 3.65585864e-01 -5.09478211e-01
-9.99051690e-01 -7.76014328e-01 7.97457099e-01 3.24157387e-01
-3.60273689e-01 8.84597421e-01 5.17660141e-01 -6.39429927e-01
-1.30676493e-01 -8.99503767e-01 -7.15984106e-01 -1.81089491e-01
4.54472721e-01 8.81375134e-01 6.47874832e-01 -5.71294188e-01
5.20690799e-01 9.00743380e-02 -1.53876805e+00 5.76572418e-01
1.30678833e+00 6.49809599e-01 6.52031243e-01 1.10524786e+00
-7.02818096e-01 1.23775756e+00 -9.68947113e-01 -2.72529721e-01
-2.79273719e-01 -1.06764627e+00 -1.75195038e+00 1.93776488e-01
1.58828467e-01 -5.06297767e-01 1.14720827e-03 6.57480597e-01
2.05593050e-01 4.57683891e-01 8.95552456e-01 -1.00782061e+00
-1.08110023e+00 6.58037841e-01 3.44302624e-01 2.42010772e-01
9.69678879e-01 -3.78214628e-01 6.05726123e-01 -6.96247518e-01
5.40609777e-01 1.31358397e+00 1.63103908e-01 9.75186288e-01
-8.50540221e-01 -2.01189414e-01 1.06923737e-01 4.10777509e-01
-8.70290399e-01 -5.03671825e-01 7.89145291e-01 -3.90537918e-01
1.01806736e+00 1.90494031e-01 5.86451255e-02 1.11254525e+00
-9.92797017e-02 8.11310351e-01 1.06455421e+00 -1.12331474e+00
4.27209437e-01 5.69955230e-01 6.65165424e-01 2.26759598e-01
-1.94495663e-01 -3.43902200e-01 -4.83167499e-01 -2.00505242e-01
4.10329551e-01 -4.98968571e-01 -3.81715775e-01 4.51735966e-02
-1.04510641e+00 8.76078963e-01 -1.69685885e-01 5.99879384e-01
-3.59265149e-01 -5.43676972e-01 1.11547761e-01 7.64563859e-01
2.81360716e-01 7.29508162e-01 -7.54766941e-01 -1.18409298e-01
-2.56390184e-01 2.17330307e-01 1.60377514e+00 1.25486219e+00
4.79582608e-01 -6.00591481e-01 -2.58718222e-01 9.49052870e-01
1.46755427e-01 6.46971047e-01 8.30796719e-01 -1.10417914e+00
6.75400376e-01 6.62461340e-01 3.25429559e-01 -7.12189853e-01
-5.65354750e-02 5.11491075e-02 -2.66336948e-01 -5.52591205e-01
6.33980155e-01 -2.89408565e-01 -5.75145900e-01 1.60694206e+00
2.61303365e-01 -2.06231445e-01 8.28677475e-01 5.57036996e-01
1.30896950e+00 9.77414370e-01 -3.42027284e-02 -3.20197821e-01
1.45219362e+00 -1.08028436e+00 -8.77188563e-01 -2.86244571e-01
6.32184207e-01 -7.75038660e-01 1.18272924e+00 3.01998198e-01
-1.09989369e+00 -7.08033204e-01 -6.37273908e-01 -5.10914445e-01
-5.25838494e-01 -1.18676752e-01 1.20125227e-01 4.57714707e-01
-1.05673027e+00 -6.69355690e-02 -9.99117717e-02 -9.20991659e-01
-2.30276436e-01 -2.80429795e-02 -9.19571891e-02 -5.54463387e-01
-1.41855860e+00 9.98149276e-01 4.30274934e-01 -2.35255077e-01
-6.98234499e-01 -3.76442045e-01 -8.68342161e-01 9.61098224e-02
6.42048180e-01 -7.55533695e-01 1.96288323e+00 -1.20101249e+00
-1.92759597e+00 6.97073400e-01 -6.38424039e-01 -3.11353445e-01
-1.18089698e-01 -1.79702148e-01 -4.13307786e-01 4.44016963e-01
2.57474184e-01 3.73328745e-01 6.96357667e-01 -1.36556780e+00
-7.43445456e-01 -3.32770675e-01 5.48633099e-01 3.86423081e-01
-3.04824829e-01 8.57524723e-02 -4.01275456e-01 -3.42820942e-01
-2.67992815e-04 -5.28707147e-01 2.64442384e-01 -6.30605578e-01
-7.41029829e-02 -7.38839269e-01 6.20521605e-01 -1.05471432e+00
1.10377562e+00 -1.66230583e+00 1.82600006e-01 8.28422457e-02
-1.30686210e-02 3.19670707e-01 -4.28125560e-01 8.59937608e-01
1.10019431e-01 -1.27174733e-02 -3.47480953e-01 2.87292823e-02
4.72411849e-02 2.88604528e-01 -5.24073660e-01 -4.78444159e-01
2.17872947e-01 8.54907811e-01 -1.11947644e+00 -4.97537583e-01
-1.90139189e-01 -4.76071658e-03 -4.69097316e-01 8.61017585e-01
-6.48618639e-01 6.23678207e-01 -7.78096497e-01 2.14798406e-01
4.26124334e-01 -2.40209848e-01 1.66685328e-01 3.72202545e-01
1.07297115e-01 6.83710575e-01 -8.79331648e-01 1.50125480e+00
-5.07975101e-01 5.24444044e-01 -3.39078009e-02 -1.17350483e+00
1.07505822e+00 6.10345185e-01 -1.02989256e-01 -9.46154714e-01
4.10475768e-02 9.17821527e-02 -7.69912377e-02 -1.11605906e+00
5.74783385e-01 -2.45774940e-01 -1.94980338e-01 9.71717477e-01
5.15152961e-02 -4.04751390e-01 4.25451636e-01 4.52453285e-01
1.19813228e+00 -1.72379091e-01 3.34107578e-01 2.54160259e-02
1.05952954e+00 3.65279853e-01 -2.42157392e-02 9.07294154e-01
6.59897402e-02 3.19850564e-01 4.61149246e-01 1.49792090e-01
-6.01996601e-01 -1.01396585e+00 1.79345801e-01 1.49879420e+00
-1.54069483e-01 -1.65709868e-01 -1.10139537e+00 -8.30396950e-01
-3.29271853e-01 1.34447217e+00 -3.86020988e-01 -2.13933408e-01
-6.99362874e-01 1.61248744e-01 3.83753061e-01 4.89056259e-01
7.15247631e-01 -1.34872973e+00 -3.43537480e-01 5.16886413e-01
-9.19406772e-01 -1.18867278e+00 -4.22098696e-01 -9.84713882e-02
-7.98349559e-01 -1.10972059e+00 -7.77475476e-01 -1.33224976e+00
9.57716763e-01 3.63165379e-01 1.32683992e+00 2.68868059e-01
4.88504380e-01 1.07606876e+00 -9.67506528e-01 -7.85144150e-01
-9.89279628e-01 2.08095834e-01 -3.81147236e-01 -1.66332692e-01
8.88128102e-01 -2.65900195e-01 -1.93649843e-01 4.61534232e-01
-1.26690543e+00 6.93833455e-02 4.66961116e-01 4.71169472e-01
3.18675563e-02 1.61319315e-01 1.08181536e+00 -8.73145401e-01
1.41877592e+00 -6.29063845e-01 -1.55766875e-01 6.38681114e-01
-9.77228507e-02 1.02464959e-01 6.72076046e-01 -4.05957699e-01
-1.53179562e+00 -3.73003274e-01 -2.73677349e-01 4.40559357e-01
-6.06247008e-01 5.61656475e-01 -3.56600881e-01 4.68385279e-01
9.25192177e-01 4.83669192e-01 -1.91379804e-02 -6.13731742e-01
2.62782186e-01 1.27607620e+00 5.53939044e-01 -9.84075427e-01
6.42396331e-01 -1.75864905e-01 -7.01639354e-01 -1.00266767e+00
-1.05441093e+00 -7.80472517e-01 -8.78119648e-01 -1.99480787e-01
8.51773620e-01 -5.50244808e-01 -3.65470856e-01 3.67775053e-01
-1.34434068e+00 -4.86642420e-01 -2.77903885e-01 2.28654012e-01
-5.54777682e-01 4.69624698e-01 -3.22715729e-01 -7.97629893e-01
-2.80336589e-01 -6.16814673e-01 8.57113719e-01 3.38852793e-01
-5.01269877e-01 -1.08677673e+00 2.48771638e-01 9.39431608e-01
4.46947038e-01 -2.35253364e-01 1.33529675e+00 -1.14413285e+00
-2.23816589e-01 -9.13629755e-02 -4.72729057e-02 7.16667652e-01
4.10615623e-01 -4.98328209e-01 -8.85630548e-01 -1.81474183e-02
4.77070868e-01 -6.37891352e-01 2.67010629e-01 -3.92741077e-02
8.85346234e-01 -3.84892374e-01 -1.69962853e-01 -4.51294154e-01
8.59410524e-01 6.27238631e-01 6.82614982e-01 3.65172297e-01
1.08916394e-01 1.14779806e+00 6.42562807e-01 -2.17511162e-01
1.01683819e+00 3.74388516e-01 -3.41562003e-01 2.99929261e-01
-7.59477541e-02 -3.74785751e-01 4.15297568e-01 1.23457921e+00
2.05660239e-01 -3.97774696e-01 -1.11013329e+00 9.18252051e-01
-1.69759488e+00 -8.52424264e-01 -5.93571588e-02 1.84292853e+00
1.10442007e+00 -1.67064071e-01 -1.41703427e-01 4.74212281e-02
5.75370073e-01 -4.57633793e-01 -4.54993516e-01 -6.48229778e-01
2.34540269e-01 3.13694745e-01 -3.98106039e-01 7.27067888e-01
-5.71891606e-01 8.58394980e-01 6.07754040e+00 1.32357359e-01
-4.84226286e-01 3.44289467e-02 9.85549837e-02 5.81389189e-01
-3.69784653e-01 1.17537707e-01 -6.74265563e-01 1.44246072e-01
1.23802233e+00 -4.31149274e-01 3.19104582e-01 6.98276639e-01
1.67555183e-01 -3.14033598e-01 -1.34201097e+00 2.85270035e-01
6.51267886e-01 -7.38975406e-01 3.41451913e-01 -2.96370924e-01
5.30594528e-01 -6.63527966e-01 -4.71916825e-01 7.40570307e-01
2.82770991e-01 -9.16861653e-01 2.83311963e-01 7.31690407e-01
1.69161484e-01 -6.54688954e-01 8.70428324e-01 9.84059095e-01
-6.14823222e-01 -1.06576219e-01 -3.78298283e-01 -1.47776380e-01
1.79648414e-01 1.56400412e-01 -1.22564483e+00 4.81807381e-01
5.73788643e-01 1.68419361e-01 -5.33947647e-01 7.01598704e-01
-4.77761835e-01 7.30734289e-01 1.67057700e-02 -3.08795094e-01
4.19489341e-03 -3.24435346e-02 3.97918165e-01 1.02002311e+00
2.28116482e-01 6.63459957e-01 4.03217738e-03 5.86123765e-01
-1.64262280e-01 3.37811291e-01 -6.20111585e-01 -1.58545583e-01
4.43854302e-01 6.91747904e-01 -3.27387512e-01 -5.77270091e-01
-5.35664618e-01 1.02438760e+00 2.84849733e-01 4.96280491e-01
-1.79240018e-01 -8.08593094e-01 -8.04310739e-02 1.34449944e-01
2.25015387e-01 -1.50450751e-01 2.42324676e-02 -1.13610065e+00
2.67500490e-01 -1.45131028e+00 4.32335138e-01 -1.21058238e+00
-1.47203696e+00 5.83052099e-01 3.92557502e-01 -8.73867571e-01
-7.50096560e-01 -4.08366203e-01 -5.59511781e-01 1.08178067e+00
-1.61460125e+00 -7.30335891e-01 -4.78009015e-01 8.48946154e-01
9.75732863e-01 -2.01890916e-01 1.11004245e+00 -1.01712629e-01
-7.84180760e-02 2.30585784e-01 8.37437585e-02 2.10382581e-01
8.13764036e-01 -1.28502452e+00 2.45486170e-01 6.75270379e-01
-1.90174684e-01 8.90125334e-01 7.74400592e-01 -6.26941442e-01
-1.23609149e+00 -7.58963287e-01 1.64813125e+00 -8.01218331e-01
5.19366086e-01 -3.71215641e-01 -1.46340430e+00 8.54652166e-01
7.06148148e-01 -1.01320326e+00 9.95663643e-01 -1.28912956e-01
-1.91552341e-02 1.69815585e-01 -1.20657218e+00 5.57605982e-01
7.18306422e-01 -6.63297296e-01 -1.84633255e+00 5.42793155e-01
8.78387213e-01 -2.93335497e-01 -8.62652004e-01 -2.77696131e-03
6.46606013e-02 -5.04410625e-01 6.76415741e-01 -6.85958505e-01
4.98756766e-01 -2.48462915e-01 -2.03698054e-01 -1.40347219e+00
7.12379217e-02 -3.50785464e-01 -1.16538607e-01 1.50152719e+00
5.80412626e-01 -7.26026237e-01 1.82017908e-01 9.58000422e-01
-1.60280466e-01 -7.64731541e-02 -6.19396091e-01 -5.63548923e-01
4.91579056e-01 -4.07426476e-01 7.23124206e-01 7.37861216e-01
3.52686435e-01 9.07406211e-01 2.34687284e-01 2.26338699e-01
1.96887508e-01 1.43353075e-01 1.07972729e+00 -1.37628019e+00
-1.98981151e-01 7.47130290e-02 5.42068243e-01 -1.68245208e+00
3.90419215e-01 -7.21894264e-01 5.63255489e-01 -2.04384828e+00
8.15701038e-02 2.52903607e-02 2.64328331e-01 2.23499194e-01
-2.91905284e-01 -6.19694948e-01 -2.19155233e-02 2.70955265e-01
-5.34734607e-01 4.65907514e-01 1.46678472e+00 -3.61602800e-03
-3.59650582e-01 2.26140171e-01 -1.30242264e+00 6.41428888e-01
9.48067427e-01 -4.88274902e-01 -9.87807095e-01 -6.31271124e-01
-4.39014798e-03 2.38059968e-01 6.98605552e-02 -7.50994027e-01
5.42465448e-01 -1.65782213e-01 1.21112853e-01 -5.59961498e-01
5.58167733e-02 -7.35742629e-01 -5.01620114e-01 6.88550668e-03
-7.10669518e-01 -1.00882892e-02 3.37301224e-01 4.21461105e-01
-4.51424599e-01 -8.76321554e-01 3.71174544e-01 -3.32300305e-01
-7.46784449e-01 -4.92287576e-01 -9.96262550e-01 3.65908295e-01
8.17993164e-01 -2.08519146e-01 -5.38018942e-01 -9.33701694e-01
-5.30621707e-01 4.63026345e-01 1.93111792e-01 7.31189549e-01
6.58799350e-01 -9.79602635e-01 -8.27764988e-01 9.63327512e-02
2.89852798e-01 2.16179162e-01 3.26802246e-02 2.37688959e-01
-2.80328274e-01 9.67993021e-01 5.20479344e-02 -4.11719561e-01
-1.12628257e+00 4.47268307e-01 1.33317053e-01 -3.78941327e-01
-4.27391857e-01 4.56149817e-01 1.54721871e-01 -8.95876527e-01
2.39583626e-01 -2.82856524e-01 -9.29966807e-01 3.48306298e-02
9.62170482e-01 1.39480323e-01 1.23186067e-01 -4.28541958e-01
1.03495084e-02 4.50403780e-01 -3.38782638e-01 -4.63374048e-01
1.03370917e+00 -5.43039978e-01 -1.94702402e-01 3.85600179e-01
1.02119911e+00 -4.70055863e-02 -6.59279823e-01 -6.26804113e-01
2.38883272e-01 -1.10017411e-01 -3.76348615e-01 -1.05964470e+00
-1.30313784e-01 7.02270746e-01 3.25309448e-02 4.95155245e-01
1.18859148e+00 3.98109615e-01 9.13058877e-01 1.03955519e+00
3.52791339e-01 -1.33051920e+00 4.58357871e-01 1.17012537e+00
1.34803092e+00 -1.04754174e+00 -5.12466609e-01 -2.91453183e-01
-8.08910012e-01 1.17496419e+00 7.99658060e-01 4.12640005e-01
4.85937268e-01 -2.37530261e-01 3.78707021e-01 -2.24191070e-01
-8.62488329e-01 -2.73821592e-01 3.63131583e-01 9.93599474e-01
4.14627820e-01 -4.82729554e-01 -3.07647377e-01 8.51311862e-01
-5.32416224e-01 3.07171583e-01 6.98784709e-01 1.46831751e+00
-7.69400895e-01 -1.27457690e+00 -6.10428452e-01 2.68922746e-01
-2.58588195e-01 -9.78647768e-02 -9.44018483e-01 7.52725720e-01
-3.26295257e-01 1.86160529e+00 -8.65028799e-03 -2.77787969e-02
9.25617158e-01 5.85208714e-01 2.75486022e-01 -1.11313832e+00
-5.88627338e-01 -5.17720759e-01 3.89620572e-01 4.57067378e-02
-6.08489633e-01 -4.27230954e-01 -1.22192013e+00 -3.09053585e-02
-1.74906790e-01 8.47741842e-01 5.35806835e-01 1.38604021e+00
2.40518153e-01 3.86454076e-01 4.31873053e-01 -1.17452994e-01
-4.60097224e-01 -1.30984962e+00 -1.33369640e-01 3.83640498e-01
2.77369320e-01 -1.06463872e-01 -4.66982961e-01 5.03543556e-01]
|
[11.732295036315918, 8.015398979187012]
|
9611b7ad-db53-4f2c-b856-f7476eba5a48
|
real-time-mapping-of-tissue-properties-for
|
2107.08120
| null |
https://arxiv.org/abs/2107.08120v1
|
https://arxiv.org/pdf/2107.08120v1.pdf
|
Real-Time Mapping of Tissue Properties for Magnetic Resonance Fingerprinting
|
Magnetic resonance Fingerprinting (MRF) is a relatively new multi-parametric quantitative imaging method that involves a two-step process: (i) reconstructing a series of time frames from highly-undersampled non-Cartesian spiral k-space data and (ii) pattern matching using the time frames to infer tissue properties (e.g., T1 and T2 relaxation times). In this paper, we introduce a novel end-to-end deep learning framework to seamlessly map the tissue properties directly from spiral k-space MRF data, thereby avoiding time-consuming processing such as the nonuniform fast Fourier transform (NUFFT) and the dictionary-based Fingerprint matching. Our method directly consumes the non-Cartesian k- space data, performs adaptive density compensation, and predicts multiple tissue property maps in one forward pass. Experiments on both 2D and 3D MRF data demonstrate that quantification accuracy comparable to state-of-the-art methods can be accomplished within 0.5 second, which is 1100 to 7700 times faster than the original MRF framework. The proposed method is thus promising for facilitating the adoption of MRF in clinical settings.
|
['Pew-Thian Yap', 'Yong Chen', 'Yilin Liu']
|
2021-07-16
| null | null | null | null |
['magnetic-resonance-fingerprinting']
|
['medical']
|
[ 3.74580592e-01 -2.81312346e-01 -5.23933023e-02 -4.89739925e-01
-1.16881156e+00 -3.22352767e-01 2.72050470e-01 1.21184587e-01
-6.84090197e-01 6.20239973e-01 1.50697947e-01 -1.60375968e-01
-4.67358321e-01 -4.06996459e-01 -5.98112404e-01 -9.69672263e-01
-4.30377275e-01 6.91947281e-01 4.10574377e-01 4.92070287e-01
1.79934502e-01 5.80372453e-01 -9.88282681e-01 1.79987431e-01
6.74922347e-01 1.16243446e+00 4.56605077e-01 6.51459396e-01
1.99949354e-01 8.92448127e-01 -3.63918790e-03 1.87459975e-01
-4.77352040e-03 -2.12213844e-01 -1.01608539e+00 -2.75668204e-01
5.12492120e-01 -7.98678517e-01 -7.04302430e-01 7.81186104e-01
8.41678977e-01 2.64426172e-01 6.70060158e-01 -5.47986269e-01
-4.39002752e-01 4.13753271e-01 -7.61425138e-01 6.69384241e-01
-1.13983899e-01 4.38142149e-03 1.75897658e-01 -7.93895662e-01
6.02243245e-01 5.99478066e-01 9.58175480e-01 3.17540586e-01
-1.39433086e+00 -5.25352299e-01 -5.67019343e-01 3.55143286e-02
-1.38999987e+00 -3.83146703e-01 6.87805057e-01 -7.92937458e-01
7.80759215e-01 8.19140673e-02 5.84632993e-01 6.25834763e-01
7.80652165e-01 5.13058364e-01 1.37196028e+00 -1.36865780e-01
2.05084793e-02 -6.01520240e-01 -6.85636476e-02 6.96265936e-01
-1.84276596e-01 1.93788067e-01 -4.00282592e-01 -5.01194656e-01
1.28642988e+00 -1.20897077e-01 -4.64532912e-01 -3.35522592e-01
-1.93503106e+00 4.90514487e-01 2.14229271e-01 4.80999827e-01
-7.41195083e-01 4.30346668e-01 6.41632915e-01 -1.51926875e-01
5.67085624e-01 2.20102087e-01 -2.49732673e-01 -2.44064301e-01
-1.36287272e+00 2.20318601e-01 1.18475899e-01 4.47720140e-01
3.85436237e-01 -1.65322617e-01 -2.93175936e-01 9.13486958e-01
2.69816190e-01 6.22035563e-01 8.35241497e-01 -1.35585761e+00
-1.42459884e-01 -2.76491165e-01 1.44675538e-01 -8.66192102e-01
-7.25359559e-01 -3.15958977e-01 -9.15438235e-01 -2.78558522e-01
7.54011512e-01 9.48100090e-02 -8.33052874e-01 1.60351121e+00
6.79109633e-01 6.65894389e-01 -5.64844131e-01 1.44133115e+00
7.96638787e-01 3.81426871e-01 1.57030299e-02 -4.49707597e-01
1.33144164e+00 -6.45760179e-01 -7.67695546e-01 2.21480668e-01
6.06959939e-01 -6.42542660e-01 8.65972877e-01 2.33626425e-01
-1.14753199e+00 -4.06961054e-01 -8.06346476e-01 -1.46214291e-01
2.45582223e-01 8.04221109e-02 7.89561689e-01 3.96606386e-01
-8.94503593e-01 8.78347993e-01 -1.31489956e+00 2.09687248e-01
4.93776619e-01 5.61745048e-01 -5.32237113e-01 -3.67313206e-01
-1.20285511e+00 7.93555140e-01 -1.31985238e-02 1.90027028e-01
-9.86057937e-01 -1.65809548e+00 -6.09226048e-01 -5.67861378e-01
-3.95410918e-02 -6.29938483e-01 1.05564046e+00 -1.59418672e-01
-1.50228357e+00 9.72628534e-01 -2.56746143e-01 -2.57026821e-01
5.89956343e-01 7.21449479e-02 -3.70089740e-01 6.48812175e-01
3.85122865e-01 4.28546339e-01 6.07886016e-01 -8.79956663e-01
8.40933695e-02 -6.35585964e-01 -4.99401867e-01 5.43530658e-02
1.85844555e-01 9.45526585e-02 -1.77797586e-01 -6.85354888e-01
4.82373863e-01 -9.32492077e-01 -2.46437758e-01 4.71126735e-01
-2.11677611e-01 3.95650834e-01 3.12854081e-01 -1.02430236e+00
9.31025028e-01 -1.84421384e+00 -2.55650729e-02 1.49600282e-01
7.21201777e-01 -6.88591674e-02 -1.93708576e-02 -2.16777131e-01
-2.46104568e-01 -2.96419621e-01 -4.54404712e-01 -3.75597626e-02
-4.76624250e-01 -1.73391879e-01 4.29676548e-02 1.19132757e+00
-3.44461262e-01 9.98388946e-01 -1.24079418e+00 -6.77866340e-01
3.17271918e-01 7.61144936e-01 -2.86328971e-01 1.17979720e-01
4.05715436e-01 1.20316744e+00 -3.71016771e-01 6.08536959e-01
7.67474771e-01 -3.12043458e-01 2.54817724e-01 -7.86292553e-01
-3.02420169e-01 2.81848669e-01 -5.87557673e-01 2.17449093e+00
-4.36788738e-01 4.47548121e-01 1.40689343e-01 -1.17408848e+00
7.51321733e-01 5.28355062e-01 1.43698287e+00 -9.53480840e-01
1.00560986e-01 5.11680901e-01 7.34034984e-04 -6.51353419e-01
9.81666297e-02 -7.11065590e-01 2.61810213e-01 6.74182534e-01
1.47241607e-01 -1.79135352e-01 -1.22474104e-01 -2.87411749e-01
1.04151058e+00 1.09879985e-01 -1.21468499e-01 -6.77588522e-01
4.59391475e-01 -2.29814157e-01 4.94831592e-01 5.51105499e-01
-5.86543798e-01 5.99671304e-01 1.77980736e-01 -7.57700562e-01
-1.33920789e+00 -1.28866279e+00 -6.85328305e-01 5.77410638e-01
8.68007392e-02 -4.41830186e-03 -6.01599872e-01 -2.95151055e-01
1.87751725e-01 -6.93696411e-03 -8.57732356e-01 7.27585256e-02
-8.50173175e-01 -1.14995432e+00 7.07772374e-01 4.42261428e-01
3.18674296e-01 -5.68468451e-01 -7.00695395e-01 6.33025110e-01
-5.65790057e-01 -1.21738207e+00 -7.34280586e-01 8.31169635e-02
-1.27248991e+00 -7.50284612e-01 -1.24705184e+00 -5.15383542e-01
4.19951677e-01 2.61459857e-01 9.24892366e-01 -2.12318525e-01
-5.08558452e-01 4.48535718e-02 1.07992198e-02 1.29942998e-01
-2.44318441e-01 -1.61723658e-01 1.58957049e-01 -3.24870367e-03
9.63456333e-02 -8.17599535e-01 -1.03428543e+00 3.47896546e-01
-6.60205483e-01 1.67481557e-01 6.01405501e-01 8.38517606e-01
1.43136823e+00 -1.04676180e-01 5.65474510e-01 -6.97131157e-01
2.02732161e-01 -4.00761783e-01 -3.56076509e-01 2.29169458e-01
-3.87791246e-01 1.79727264e-02 3.81002903e-01 -6.90408528e-01
-7.33864009e-01 8.49051103e-02 -6.51568845e-02 -6.37182415e-01
7.67762512e-02 6.15760982e-01 5.55080712e-01 -5.26216328e-01
5.73241711e-01 7.29281306e-01 5.19759059e-01 -3.18575084e-01
9.71254054e-03 3.69199485e-01 1.02063918e+00 -8.44185412e-01
3.14173102e-01 8.46746504e-01 3.71026933e-01 -8.30995321e-01
-6.06207550e-01 -5.33014119e-01 -9.78853226e-01 -6.51279986e-01
9.06820476e-01 -6.98480964e-01 -6.86812878e-01 7.97631502e-01
-8.72445762e-01 -5.85806251e-01 -4.94116843e-02 1.03218853e+00
-8.32421124e-01 6.96063221e-01 -1.03512156e+00 -3.45756590e-01
-6.92798376e-01 -1.35720193e+00 1.24477553e+00 -9.73862037e-02
-1.21894605e-01 -1.05076241e+00 3.61844838e-01 4.94293183e-01
6.83283031e-01 5.68939507e-01 8.25091422e-01 7.82034453e-03
-2.07734287e-01 -1.89301185e-02 -1.69889286e-01 -2.48032738e-03
1.63444743e-01 -4.69354570e-01 -7.39665151e-01 -2.89965510e-01
2.39711091e-01 -4.07073021e-01 5.56443274e-01 1.15052855e+00
1.53731143e+00 2.35410452e-01 -1.48956254e-01 1.08595586e+00
1.24762082e+00 1.67899787e-01 7.23141849e-01 1.39107659e-01
7.24341750e-01 4.31250602e-01 6.26045108e-01 3.88266116e-01
4.18254226e-01 7.91214168e-01 1.13488715e-02 -2.44757831e-01
-1.97589740e-01 -1.30128458e-01 -2.11514622e-01 1.05721927e+00
-1.52568936e-01 6.32508934e-01 -1.06431270e+00 5.55730283e-01
-1.46796227e+00 -8.90626490e-01 -2.68558174e-01 2.19419098e+00
1.08618104e+00 -4.35152173e-01 1.18189454e-01 2.92660445e-02
6.51820123e-01 1.05054341e-01 -9.29127038e-01 4.11912441e-01
3.03218693e-01 3.99049342e-01 7.37358987e-01 5.57102203e-01
-1.12011921e+00 3.46045375e-01 6.62591505e+00 6.85262024e-01
-1.67877746e+00 4.74836469e-01 8.83131504e-01 -1.75077468e-01
-3.49540114e-01 -4.29150701e-01 -5.13126925e-02 5.63111126e-01
1.13919938e+00 2.53300834e-02 7.28522956e-01 3.06350440e-01
4.50316548e-01 -1.04998320e-01 -7.05899060e-01 1.28367877e+00
-3.43615502e-01 -1.49538445e+00 -4.16539162e-01 1.35277003e-01
2.56745398e-01 4.42212701e-01 -9.13290158e-02 -2.88950235e-01
-4.37771916e-01 -1.08387601e+00 7.18738675e-01 8.11878502e-01
1.48197830e+00 -4.30328965e-01 4.42776233e-01 1.19099401e-01
-1.03585625e+00 4.66307968e-01 -3.92444164e-01 5.37911355e-01
3.15848827e-01 1.23095322e+00 -6.61178648e-01 4.74627376e-01
6.60229087e-01 7.16157317e-01 -7.89871514e-02 1.13170421e+00
4.79833066e-01 4.47063833e-01 -2.49512330e-01 5.28749406e-01
-6.12294674e-03 -1.93656191e-01 3.38149965e-01 1.06311524e+00
3.90980363e-01 3.60929906e-01 -1.07433304e-01 9.25447106e-01
3.02437901e-01 -1.21472031e-01 -3.33311819e-02 -3.64473031e-04
4.92798120e-01 1.37688947e+00 -8.88730824e-01 -2.75564134e-01
-2.89059818e-01 7.22437739e-01 1.73046310e-02 1.13656841e-01
-1.02077699e+00 -2.12319165e-01 2.30206624e-01 6.83259726e-01
-1.78788286e-02 -3.93249124e-01 -2.85692364e-02 -1.09894407e+00
-1.36442378e-01 -4.57812607e-01 4.64772135e-02 -8.51701975e-01
-1.31895280e+00 5.85824311e-01 -1.46687385e-02 -1.04315376e+00
-9.50093344e-02 -4.76615548e-01 -1.60856262e-01 1.19964910e+00
-1.52129257e+00 -8.88485134e-01 -3.09896708e-01 6.73958004e-01
-1.37559250e-01 3.24785233e-01 8.38222682e-01 8.02021027e-01
-1.05833761e-01 4.10195976e-01 3.66157681e-01 1.01559177e-01
7.71313190e-01 -1.16057980e+00 2.02996805e-01 3.51635039e-01
-3.85015517e-01 5.94571114e-01 3.88788491e-01 -5.03508925e-01
-1.62110054e+00 -9.12878811e-01 5.58505893e-01 -1.50915548e-01
6.94462419e-01 -8.92167613e-02 -1.02484155e+00 2.13325888e-01
-5.58097005e-01 9.02980089e-01 1.01838744e+00 -4.94636029e-01
-5.03138751e-02 -2.25358874e-01 -1.51748788e+00 -1.14130892e-01
7.15177536e-01 -7.72417188e-01 -1.01672247e-01 3.17545176e-01
4.03823346e-01 -9.53006983e-01 -1.79604912e+00 4.02405500e-01
1.09295547e+00 -6.22407317e-01 1.18911684e+00 -2.00093016e-01
3.51869643e-01 -4.02049124e-01 -1.74199373e-01 -1.07074893e+00
-5.39712369e-01 -5.16286016e-01 -2.84339994e-01 4.64572489e-01
-1.10426717e-01 -5.68602562e-01 8.18116903e-01 5.85402250e-01
-3.97560745e-01 -1.06769121e+00 -1.41154945e+00 -5.14864683e-01
3.58714521e-01 -3.77164513e-01 4.65874404e-01 1.18489385e+00
-1.65874138e-01 -2.50984043e-01 -2.94320166e-01 9.44426656e-03
1.14782226e+00 9.94050056e-02 7.72313476e-02 -1.03687501e+00
-3.77995193e-01 -3.59909832e-02 -2.49425918e-01 -1.17689133e+00
3.67935188e-02 -1.03157938e+00 1.15790159e-01 -1.10167325e+00
4.67493683e-01 -8.25599968e-01 -5.00121117e-01 1.37053281e-01
2.75177695e-02 4.11525786e-01 -3.78485948e-01 5.30726910e-01
-1.68927774e-01 4.01774764e-01 2.03998017e+00 -1.02977559e-01
2.07041800e-01 -4.85467374e-01 -3.16673845e-01 1.92033648e-01
3.92381638e-01 -4.65955317e-01 -4.46241915e-01 -4.89626765e-01
-3.06792200e-01 9.22439516e-01 5.08162439e-01 -9.49757278e-01
1.18964456e-01 -7.78708011e-02 6.34015203e-01 -6.10857487e-01
1.72400862e-01 -5.13356268e-01 3.85527641e-01 5.97061932e-01
-2.79360801e-01 -4.99754809e-02 1.97056711e-01 2.51911312e-01
-8.42810273e-02 8.28173384e-02 1.11696124e+00 -1.79901123e-01
-1.85676605e-01 8.92119110e-01 -3.14473271e-01 1.31509483e-01
5.61181605e-01 -1.92343950e-01 -7.41897523e-02 5.00660315e-02
-8.82800281e-01 -2.02758789e-01 2.00414389e-01 -6.97043911e-02
6.20565653e-01 -1.56144214e+00 -6.59355581e-01 1.09534092e-01
-2.39561722e-01 9.59166363e-02 1.03822577e+00 1.57688963e+00
-8.21831286e-01 5.21965086e-01 -4.37450856e-01 -1.03746700e+00
-6.20203614e-01 3.15833211e-01 1.00341415e+00 -5.09094894e-01
-9.54959571e-01 7.62452602e-01 1.64598927e-01 -6.83656275e-01
-4.85320926e-01 -4.65283006e-01 2.14620739e-01 -2.90204018e-01
8.27079415e-01 3.37468565e-01 3.99613649e-01 -9.43806052e-01
-6.75694644e-01 7.98184991e-01 -6.14259094e-02 -2.67706335e-01
1.44246376e+00 -1.59827664e-01 -2.12819919e-01 4.22320455e-01
1.53672755e+00 -4.30983633e-01 -1.54291809e+00 -3.23253095e-01
-3.11384737e-01 -6.69621587e-01 7.98077106e-01 -8.49738955e-01
-1.39794838e+00 8.44583452e-01 7.93598652e-01 -3.72137010e-01
9.28815305e-01 -6.01680987e-02 1.27059507e+00 -3.33425969e-01
6.59496963e-01 -6.49719834e-01 -2.25247875e-01 1.24289297e-01
6.80589736e-01 -8.59901488e-01 7.56290853e-02 -4.14442062e-01
-2.38569468e-01 1.24699450e+00 1.30287126e-01 1.61108095e-02
7.66286135e-01 5.11579037e-01 4.36145216e-02 -3.64524484e-01
-4.41896558e-01 6.41244590e-01 3.38382900e-01 7.41765320e-01
8.84442806e-01 1.94812208e-01 -1.92417249e-01 4.91956919e-01
4.87738922e-02 4.40094978e-01 2.66623527e-01 7.41844833e-01
-2.23847106e-02 -7.91465163e-01 -3.14674020e-01 7.70207942e-01
-7.98528790e-01 2.87080482e-02 4.78491157e-01 3.73112530e-01
-8.24677497e-02 3.15061152e-01 4.27927859e-02 -2.10081518e-01
1.09235346e-01 -2.82032847e-01 1.18176377e+00 -2.14458510e-01
-3.10670435e-01 2.38639638e-01 -3.40412259e-01 -8.57543230e-01
-6.39128029e-01 -9.40367341e-01 -1.47564960e+00 -3.66009772e-01
-1.81114152e-02 -4.09588069e-02 6.49588227e-01 9.51300621e-01
2.45533362e-01 4.51548278e-01 5.43650150e-01 -1.16380823e+00
-3.02658349e-01 -8.15066218e-01 -8.41052234e-01 2.43243828e-01
5.05447984e-01 -8.90791595e-01 5.99717675e-03 3.76763567e-02]
|
[13.497953414916992, -2.3965036869049072]
|
c6a7e447-e6d2-4b19-a220-8c603e8c58db
|
tweetsumm-a-dialog-summarization-dataset-for-1
|
2111.11894
| null |
https://arxiv.org/abs/2111.11894v1
|
https://arxiv.org/pdf/2111.11894v1.pdf
|
TWEETSUMM -- A Dialog Summarization Dataset for Customer Service
|
In a typical customer service chat scenario, customers contact a support center to ask for help or raise complaints, and human agents try to solve the issues. In most cases, at the end of the conversation, agents are asked to write a short summary emphasizing the problem and the proposed solution, usually for the benefit of other agents that may have to deal with the same customer or issue. The goal of the present article is advancing the automation of this task. We introduce the first large scale, high quality, customer care dialog summarization dataset with close to 6500 human annotated summaries. The data is based on real-world customer support dialogs and includes both extractive and abstractive summaries. We also introduce a new unsupervised, extractive summarization method specific to dialogs.
|
['Ranit Aharonov', 'David Konopnicki', 'Sachindra Joshi', 'Benjamin Sznajder', 'Chulaka Gunasekara', 'Guy Feigenblat']
|
2021-11-23
| null | null | null | null |
['unsupervised-extractive-summarization', 'extractive-summarization']
|
['natural-language-processing', 'natural-language-processing']
|
[ 4.05776560e-01 7.93674648e-01 6.18869513e-02 -4.84547883e-01
-1.07144630e+00 -6.48230553e-01 5.17336130e-01 9.90495265e-01
-3.62259269e-01 1.16774011e+00 8.95592511e-01 1.50150478e-01
-1.98790938e-01 -3.08353841e-01 3.35473031e-01 -3.97875667e-01
6.93842351e-01 1.48907661e+00 5.65364864e-03 -4.76399988e-01
5.50154805e-01 4.26492065e-01 -8.02155793e-01 5.69454074e-01
8.78764987e-01 5.48682868e-01 5.94229758e-01 8.79063427e-01
-5.56636095e-01 9.67311561e-01 -1.25593746e+00 -4.95347440e-01
-3.38642210e-01 -7.38325179e-01 -1.26165867e+00 6.95945144e-01
-4.17276882e-02 -4.83200550e-01 1.49171486e-01 7.31024206e-01
4.72306758e-01 3.55466396e-01 6.35955334e-01 -1.14492977e+00
7.85536915e-02 8.18254411e-01 -4.44287471e-02 1.31029695e-01
7.98790812e-01 -7.73664983e-03 1.26172793e+00 -4.94200826e-01
7.88296580e-01 1.10625565e+00 3.03227782e-01 7.10111558e-01
-9.42667067e-01 2.28860974e-01 1.90191194e-01 8.86936337e-02
-5.49848974e-01 -6.35302782e-01 6.71189845e-01 -2.51751423e-01
1.13127875e+00 4.51243520e-01 3.25035423e-01 8.99982572e-01
-1.62839040e-01 7.68633485e-01 2.45113254e-01 -4.21808183e-01
4.56329554e-01 5.89856148e-01 8.28816354e-01 3.63868922e-01
6.40289634e-02 -1.10556304e+00 -3.65430087e-01 -6.66381359e-01
-3.08000133e-03 -4.81146313e-02 -4.30329978e-01 9.74528790e-02
-9.11157846e-01 1.07047832e+00 -7.50889778e-02 5.00455976e-01
-7.41423488e-01 -8.08268189e-01 5.92555583e-01 2.16884911e-01
3.52625579e-01 8.36844444e-01 -4.46891993e-01 -5.26461959e-01
-5.68053067e-01 5.81137419e-01 1.73999810e+00 9.61600602e-01
4.13690805e-01 -2.94882625e-01 -2.06894383e-01 9.29425538e-01
-3.06297634e-02 -5.66508844e-02 3.94699842e-01 -1.08395398e+00
7.18055129e-01 9.55512404e-01 4.91426915e-01 -6.97334290e-01
-6.18621528e-01 -7.22640976e-02 -5.94214261e-01 -1.97651073e-01
3.99990350e-01 -3.84623051e-01 5.03166504e-02 1.00789821e+00
3.29497367e-01 -5.37701309e-01 5.75447381e-01 7.35837579e-01
1.25078011e+00 6.39888585e-01 -3.14069748e-01 -9.98340130e-01
1.76401758e+00 -1.35009849e+00 -1.33611333e+00 -1.60051182e-01
5.75913370e-01 -8.98576021e-01 6.30350530e-01 6.02455854e-01
-1.31985962e+00 -1.40371487e-01 -5.94565928e-01 -1.76589683e-01
5.22592366e-02 1.58388406e-01 1.42757535e-01 1.60449386e-01
-5.14890790e-01 2.80605257e-01 -1.98564678e-01 -7.82852769e-01
-1.56921819e-01 1.64302260e-01 -2.17783660e-01 -4.47214320e-02
-8.22405100e-01 1.01562822e+00 2.69507885e-01 8.78799614e-03
-9.39606428e-02 -3.62771869e-01 -6.29274607e-01 4.31402445e-01
8.23838472e-01 -6.52128279e-01 1.90595973e+00 -4.96068120e-01
-1.55593097e+00 6.11096978e-01 -3.29293460e-01 -4.28816289e-01
4.41527933e-01 -3.56028348e-01 -1.38494074e-01 3.19138855e-01
2.55305946e-01 1.38650596e-01 2.47811168e-01 -1.04739404e+00
-8.25111985e-01 -4.16625232e-01 1.21813335e-01 3.78014982e-01
9.03570578e-02 2.38053977e-01 -3.10001552e-01 -3.71726453e-01
-2.52601266e-01 -7.96469986e-01 -9.40824896e-02 -6.91739023e-01
-6.40046537e-01 -5.68964779e-01 8.32464755e-01 -8.91989410e-01
1.00081348e+00 -1.93629646e+00 3.95587146e-01 -2.64740586e-01
4.49755251e-01 3.91614914e-01 9.24793482e-02 1.16694725e+00
3.19397360e-01 -2.66772002e-01 -4.16139930e-01 -7.73344219e-01
6.73295408e-02 3.96474421e-01 -3.76746595e-01 -4.22479846e-02
-1.01410579e-02 6.01439953e-01 -8.14495444e-01 -6.58186674e-01
6.26720861e-02 -6.86534420e-02 -3.71161520e-01 5.48567057e-01
-4.76816326e-01 5.28789341e-01 -6.19835556e-01 1.64677933e-01
2.44124994e-01 -2.15763777e-01 4.30402994e-01 1.00102760e-02
-1.50255365e-02 7.42835283e-01 -8.72927129e-01 1.44403803e+00
-2.40077510e-01 6.20301485e-01 5.70130885e-01 -1.03692269e+00
7.21912026e-01 4.50245887e-01 4.10757691e-01 -1.14858188e-01
2.91576505e-01 4.80728522e-02 -1.47208795e-01 -5.09740293e-01
9.79282081e-01 2.59350631e-02 -2.90124416e-01 9.15600121e-01
-7.90914893e-02 -2.30319783e-01 4.91630077e-01 6.85799599e-01
1.14797664e+00 -6.72918439e-01 8.07960570e-01 2.62484968e-01
6.66605055e-01 3.58937383e-01 4.05462831e-01 6.77079320e-01
-2.56988615e-01 5.67481995e-01 9.99557793e-01 -2.64088750e-01
-5.73119044e-01 -5.75030267e-01 3.65257889e-01 8.93069744e-01
-7.15863779e-02 -6.51280940e-01 -8.13375890e-01 -8.23743284e-01
-2.64782965e-01 1.11706877e+00 -1.88877344e-01 1.43570721e-01
-7.26128757e-01 -1.92164257e-01 8.98865312e-02 1.67383239e-01
4.62587446e-01 -1.41302478e+00 -7.62538433e-01 7.06419528e-01
-8.00733566e-01 -1.46004188e+00 -5.90629399e-01 2.46960521e-02
-7.12189674e-01 -1.13382149e+00 -4.57373887e-01 -7.44235516e-01
4.20199871e-01 4.55286130e-02 9.16208684e-01 1.19729139e-01
-1.40749708e-01 5.11633456e-01 -5.98614335e-01 -2.86796272e-01
-9.51406062e-01 4.03778970e-01 -2.00038865e-01 -7.60678649e-02
2.76119918e-01 -3.27990264e-01 -4.64975871e-02 1.87826425e-01
-6.19031370e-01 -7.02638105e-02 2.82205254e-01 7.32271314e-01
-4.38057519e-02 -5.18189184e-03 9.83367741e-01 -1.05289829e+00
1.40102696e+00 -2.73857355e-01 -1.00809522e-01 3.41691762e-01
-3.03862303e-01 -3.14111169e-03 6.16754651e-01 -2.24377260e-01
-1.19411659e+00 2.12864690e-02 -2.51773775e-01 2.72454917e-01
-4.95552450e-01 5.98621011e-01 -2.11257249e-01 8.70697021e-01
3.45243275e-01 1.06850915e-01 2.15423539e-01 -7.00996578e-01
1.50041550e-01 9.50142860e-01 6.62589192e-01 -1.02856502e-01
1.28030330e-01 -1.57534387e-02 -5.90183675e-01 -9.98355687e-01
-8.36554408e-01 -8.73436093e-01 -5.50372124e-01 -2.34052703e-01
8.87857676e-01 -2.74556905e-01 -1.20861375e+00 6.31298646e-02
-1.71939981e+00 -8.64055529e-02 -4.28799361e-01 1.93339780e-01
-5.29417396e-01 4.14509207e-01 -8.27236056e-01 -1.06781149e+00
-5.37155151e-01 -9.27366138e-01 8.56458783e-01 3.87558758e-01
-1.01506305e+00 -8.28837752e-01 2.31702775e-02 8.17505002e-01
2.55640507e-01 -4.47305441e-02 1.03051841e+00 -1.66231596e+00
4.01620790e-02 -4.24878538e-01 3.24522629e-02 6.63215071e-02
6.58182263e-01 -2.19935253e-01 -5.44819474e-01 8.15721303e-02
4.92010862e-01 -2.43759423e-01 4.32087719e-01 1.90715984e-01
2.29519695e-01 -8.43208730e-01 -2.02732235e-01 -6.74626827e-01
5.43104112e-01 3.75853598e-01 4.22660887e-01 -1.83474331e-03
2.07139477e-01 1.18309295e+00 6.71704412e-01 7.49692857e-01
5.13126671e-01 7.93478370e-01 -2.50356924e-03 4.16263103e-01
2.67524958e-01 1.50747254e-01 6.92768320e-02 9.75699008e-01
3.17379385e-01 -5.02333224e-01 -5.11018872e-01 4.41168517e-01
-2.30998945e+00 -9.01721299e-01 -1.60084948e-01 1.55668366e+00
7.22416818e-01 1.26677930e-01 4.24977422e-01 1.60046205e-01
8.34287405e-01 -9.38463584e-02 -4.95887071e-01 -6.40829563e-01
1.83133587e-01 -3.92250001e-01 -2.24302590e-01 8.50573897e-01
-5.33786297e-01 6.44141853e-01 5.53185415e+00 3.65173876e-01
-5.08563340e-01 -1.19180620e-01 4.37890559e-01 -6.76172525e-02
6.85063824e-02 -1.00087725e-01 -7.66445994e-01 1.92035630e-01
8.77561510e-01 -3.14557403e-01 3.03478301e-01 8.92233908e-01
3.33023548e-01 -6.77368820e-01 -1.34017718e+00 7.87022829e-01
3.88792247e-01 -1.36465538e+00 -1.38900921e-01 3.17206942e-02
1.61831960e-01 -4.78218824e-01 -6.46849632e-01 2.03687832e-01
2.39019439e-01 -5.76160133e-01 1.96039885e-01 4.10505354e-01
-1.17199784e-02 -7.55811155e-01 1.04655981e+00 9.51218843e-01
-6.94321454e-01 -1.58338472e-01 1.68396197e-02 -2.10350275e-01
9.51571226e-01 2.55700320e-01 -1.34238672e+00 5.49944758e-01
1.99744835e-01 3.08133602e-01 -1.39678717e-01 6.89113140e-01
-8.73582810e-02 1.10093139e-01 7.04829171e-02 -3.67074490e-01
2.23687321e-01 -4.18328702e-01 7.94125855e-01 1.13141131e+00
1.57811097e-03 8.67806971e-01 3.67627949e-01 5.55091798e-01
-9.39629078e-02 3.24801236e-01 -3.43089730e-01 -2.95740515e-01
3.05490583e-01 1.46692276e+00 -8.31223905e-01 -7.54315674e-01
-1.64130718e-01 1.15606773e+00 1.99082226e-01 9.53780785e-02
-3.05997074e-01 -4.82268512e-01 3.63399982e-01 -8.71698335e-02
1.69385701e-01 -4.90530947e-05 -1.59090400e-01 -7.75308490e-01
1.19597595e-02 -1.32224786e+00 5.27947664e-01 -7.98694611e-01
-1.16136146e+00 7.71509409e-01 -1.37315437e-01 -8.41112018e-01
-8.22111070e-01 -1.32379204e-01 -8.37374985e-01 6.75820410e-01
-9.37564194e-01 -3.38006645e-01 -2.67080843e-01 3.63941133e-01
1.29884803e+00 -3.85369092e-01 1.01108468e+00 4.58258949e-02
-5.09308755e-01 -9.93592590e-02 -2.99797058e-01 5.99260535e-03
8.27603161e-01 -1.14143610e+00 2.51984764e-02 1.35913193e-01
-1.69322804e-01 5.39821506e-01 1.18409646e+00 -8.11066628e-01
-1.10933101e+00 -5.66381335e-01 1.63811648e+00 -4.11826640e-01
4.54061061e-01 -1.35190710e-01 -1.04996336e+00 4.24593180e-01
8.01922679e-01 -1.01485503e+00 7.88227737e-01 -1.34320259e-01
3.78713369e-01 1.04208335e-01 -1.44969201e+00 5.35436690e-01
4.01118577e-01 -2.90547758e-01 -1.24439764e+00 6.97419167e-01
7.53982961e-01 -3.38285863e-01 -4.74214494e-01 -1.66696429e-01
1.37459502e-01 -7.75870919e-01 3.16484123e-01 -6.56097472e-01
3.38656992e-01 1.51436046e-01 2.83447132e-02 -1.39532745e+00
2.10660592e-01 -9.98978078e-01 1.10617712e-01 1.48262906e+00
3.87814313e-01 -6.55038774e-01 6.02947474e-01 1.21759641e+00
-3.31461906e-01 -3.85723263e-01 -7.03595698e-01 -7.36782029e-02
-6.10000193e-01 -1.00716919e-01 4.61423337e-01 6.90921545e-01
7.84573793e-01 1.10357332e+00 -3.94688100e-01 -6.48790896e-02
2.99526125e-01 2.73617834e-01 9.75780010e-01 -1.51869464e+00
-1.53240740e-01 -4.45835799e-01 2.04209477e-01 -1.19102550e+00
1.98157445e-01 -3.22845995e-01 2.53634572e-01 -2.15059924e+00
3.73367313e-03 2.37225652e-01 5.57405591e-01 1.09893493e-01
7.77125359e-03 -6.48612857e-01 2.95997888e-01 2.38236725e-01
-8.25809300e-01 4.09435987e-01 7.83489048e-01 -1.39109567e-01
-7.18258798e-01 4.88971502e-01 -9.32726622e-01 7.71793664e-01
7.99961388e-01 -3.42886120e-01 -3.22399110e-01 1.14992656e-01
-4.72667590e-02 8.73478413e-01 -1.28351599e-01 -4.61786360e-01
7.19135463e-01 -1.36687741e-01 -3.43483567e-01 -9.37549770e-01
4.86961991e-01 -8.24252605e-01 -1.62608549e-01 3.79721880e-01
-8.32035840e-01 2.75411874e-01 -1.17833652e-01 3.77800703e-01
-3.41715485e-01 -7.82352924e-01 4.39122826e-01 -1.95225373e-01
-5.29693700e-02 -3.41407567e-01 -9.07476008e-01 1.09967291e-01
7.94741988e-01 1.57797206e-02 -6.21980309e-01 -1.15174758e+00
-1.05434883e+00 5.10351419e-01 1.17662989e-01 1.58496901e-01
5.54377615e-01 -7.51160920e-01 -8.52834463e-01 -3.53032768e-01
1.89271234e-02 -4.20286283e-02 1.38817027e-01 8.09535563e-01
-2.87808359e-01 5.77403724e-01 1.19032994e-01 -2.40403756e-01
-1.64315486e+00 4.70879585e-01 -1.86183959e-01 -5.03790200e-01
-9.16018546e-01 2.71411330e-01 7.17602000e-02 -2.75415182e-01
5.01430809e-01 -3.83525014e-01 -7.23587632e-01 7.44488239e-01
9.34814870e-01 5.96303523e-01 2.76573598e-01 -6.43827081e-01
-2.33240694e-01 -4.94799204e-02 -4.23462152e-01 -3.21339786e-01
1.47531760e+00 -5.79345822e-01 -2.56271154e-01 6.22687221e-01
7.65027761e-01 2.09777325e-01 -7.13347614e-01 -4.18816745e-01
3.00592661e-01 -6.52576089e-02 -7.17590749e-01 -8.13427925e-01
-5.50107598e-01 3.81701410e-01 -1.98331326e-01 9.41815317e-01
8.21071327e-01 5.45309544e-01 1.02551496e+00 9.58044291e-01
-9.48581286e-03 -1.20247018e+00 3.73554736e-01 7.67833591e-01
1.33524382e+00 -1.08237934e+00 7.54744262e-02 -4.80447829e-01
-1.37222862e+00 1.26460290e+00 3.55066776e-01 2.86348253e-01
4.30622138e-02 2.15820089e-01 1.94267690e-01 -4.11568850e-01
-1.14579618e+00 8.39492455e-02 -1.14583775e-01 2.88998842e-01
1.70692101e-01 -2.53924400e-01 -4.61280227e-01 7.62134552e-01
-3.52497280e-01 -4.45557415e-01 1.04924953e+00 9.70355332e-01
-7.60369837e-01 -1.00369596e+00 -3.10563087e-01 4.83232558e-01
-3.05444866e-01 9.20904428e-02 -1.08879125e+00 4.42204893e-01
-6.41149938e-01 1.51724911e+00 4.59528938e-02 2.50166118e-01
7.86537051e-01 4.55888689e-01 -3.41937542e-02 -9.45974767e-01
-1.06988156e+00 2.68475950e-01 8.17182600e-01 2.92882714e-02
-4.89490032e-01 -7.97388911e-01 -1.48554921e+00 -1.67748570e-01
-2.55681247e-01 7.85969973e-01 5.17683446e-01 1.22441947e+00
2.76717365e-01 5.28838158e-01 5.09149134e-01 -8.19424689e-01
-3.95282000e-01 -1.25377524e+00 -4.96322483e-01 4.19612199e-01
4.25575167e-01 -1.98442906e-01 -2.87662536e-01 6.93572015e-02]
|
[12.676711082458496, 8.757143020629883]
|
3aa5b967-1f4b-4350-9384-63acf538bc8b
|
agile-the-first-lemmatizer-for-ancient-greek
| null | null |
https://aclanthology.org/2022.lrec-1.571
|
https://aclanthology.org/2022.lrec-1.571.pdf
|
AGILe: The First Lemmatizer for Ancient Greek Inscriptions
|
To facilitate corpus searches by classicists as well as to reduce data sparsity when training models, we focus on the automatic lemmatization of ancient Greek inscriptions, which have not received as much attention in this sense as literary text data has. We show that existing lemmatizers for ancient Greek, trained on literary data, are not performant on epigraphic data, due to major language differences between the two types of texts. We thus train the first inscription-specific lemmatizer achieving above 80% accuracy, and make both the models and the lemmatized data available to the community. We also provide a detailed error analysis highlighting peculiarities of inscriptions which again highlights the importance of a lemmatizer dedicated to inscriptions.
|
['Malvina Nissim', 'Saskia Peels-Matthey', 'Jasper K. Bos', 'Silvia Stopponi', 'Evelien de Graaf']
| null | null | null | null |
lrec-2022-6
|
['lemmatization']
|
['natural-language-processing']
|
[ 1.05822548e-01 1.68695197e-01 -9.83135700e-02 -7.30436742e-02
-9.52076614e-01 -7.98159361e-01 9.11133349e-01 3.24933767e-01
-9.94285047e-01 6.44850492e-01 7.42933273e-01 -5.75908899e-01
-1.35351539e-01 -8.49959016e-01 -4.01395649e-01 -3.11598957e-01
3.49095672e-01 8.97840381e-01 -6.32385835e-02 -4.08890933e-01
2.22338974e-01 3.48714709e-01 -1.11836076e+00 3.90855968e-01
8.05778086e-01 4.20376599e-01 -7.08567947e-02 2.65567183e-01
-3.08149487e-01 2.93691248e-01 -7.27330506e-01 -1.03450549e+00
4.42845255e-01 -2.51863062e-01 -7.96158731e-01 -2.08951384e-01
5.12449861e-01 -7.73607567e-02 -5.10435998e-01 6.78075075e-01
6.26975596e-01 -7.93884620e-02 8.49991679e-01 -3.13978374e-01
-5.83954453e-01 1.47719812e+00 -6.96792305e-02 3.85732651e-01
2.79001445e-01 -2.30513960e-01 1.48359799e+00 -9.93527412e-01
1.05912924e+00 9.32827950e-01 7.79389858e-01 3.54112744e-01
-1.34736824e+00 -4.19054985e-01 -1.39853314e-01 4.25600141e-01
-1.38202667e+00 -5.51363230e-01 8.45016360e-01 -3.61033022e-01
1.08252263e+00 2.52425849e-01 7.94056416e-01 1.22881913e+00
-8.13969895e-02 1.00072861e+00 9.96587217e-01 -6.40167058e-01
-3.82664725e-02 -1.44758001e-01 -3.16904150e-02 4.49740708e-01
1.73851579e-01 -6.39947280e-02 -8.20708036e-01 -6.90561831e-02
4.40645695e-01 -4.78628278e-01 -1.38922468e-01 8.31468552e-02
-1.26362574e+00 8.83733213e-01 -6.26025945e-02 6.30613506e-01
-2.99140126e-01 -9.16781873e-02 9.58500981e-01 5.80308616e-01
5.03009796e-01 6.29197478e-01 -4.86164153e-01 -5.68766356e-01
-1.48199189e+00 1.32417053e-01 8.83922338e-01 8.35824192e-01
5.22353411e-01 3.48662809e-02 2.84854650e-01 9.90720391e-01
-8.37105364e-02 6.53451204e-01 4.85672414e-01 -2.44977310e-01
8.87376070e-01 5.45849383e-01 -6.18591726e-01 -7.34010041e-01
-3.58833730e-01 -3.15552801e-01 -6.78220034e-01 -3.21941316e-01
4.18879807e-01 5.83338812e-02 -7.09480822e-01 1.35040140e+00
2.41440367e-02 -5.29486895e-01 -5.11089116e-02 4.89936531e-01
9.30780649e-01 4.48837370e-01 7.11897314e-02 -3.24884593e-01
1.44189286e+00 -4.39626247e-01 -6.58138096e-01 -2.80377150e-01
9.78065789e-01 -1.07657456e+00 1.32468224e+00 5.05353212e-01
-1.11734116e+00 -1.43176049e-01 -1.08204103e+00 -5.16078770e-01
-4.80708659e-01 4.18171138e-01 5.43872774e-01 6.83107197e-01
-3.88125658e-01 8.13470542e-01 -8.82901490e-01 -3.43463868e-01
5.77677608e-01 6.30043671e-02 -3.46482545e-01 3.57154071e-01
-1.22642195e+00 1.38580954e+00 5.87578773e-01 -9.92573798e-02
-3.51847976e-01 -9.18320715e-01 -7.36128211e-01 -9.57018062e-02
5.61748862e-01 -1.03852853e-01 1.00252712e+00 -6.59895420e-01
-1.24046433e+00 1.40038610e+00 2.76245564e-01 -5.92111230e-01
6.75572336e-01 -2.29874015e-01 -4.52196568e-01 -6.32883608e-02
1.19085580e-01 1.44657800e-02 4.96589363e-01 -7.55606294e-01
-6.06537580e-01 -1.46004632e-01 -2.19675109e-01 4.52833511e-02
-5.76040387e-01 2.32882082e-01 -6.00872755e-01 -1.17307866e+00
3.39126319e-01 -7.74742067e-01 1.10303037e-01 -2.45506927e-01
-6.38544023e-01 -4.21095163e-01 4.84565735e-01 -1.05681813e+00
1.82732844e+00 -1.90404892e+00 1.31441191e-01 4.17110175e-01
-7.42804334e-02 2.43909657e-01 -5.45678250e-02 8.32389295e-01
-2.34912913e-02 2.45350182e-01 -3.77677292e-01 -5.81522584e-01
3.34582686e-01 6.85494483e-01 -2.86485374e-01 7.03171313e-01
1.07645214e-01 1.00181806e+00 -7.47363687e-01 -6.10848129e-01
-1.47640472e-03 2.62136847e-01 -5.43414116e-01 -2.72910535e-01
-2.10261866e-01 -1.20611750e-01 -2.29947716e-01 6.17507517e-01
3.53193939e-01 2.81403393e-01 5.20817697e-01 -2.58482754e-01
-3.10768485e-01 1.24261165e+00 -9.79258120e-01 1.87717712e+00
-6.70551777e-01 8.40975046e-01 -2.14977726e-01 -8.55686247e-01
7.42152393e-01 3.15068483e-01 5.84255636e-01 -9.04702246e-01
3.80067438e-01 5.02775371e-01 1.16217852e-01 -4.07993287e-01
7.48617470e-01 -6.29394352e-01 -4.16681826e-01 9.55257595e-01
2.54200529e-02 -1.19085819e-01 8.16925764e-01 3.24546427e-01
1.19726121e+00 1.07094347e-01 4.71240282e-01 -4.85708207e-01
4.14115757e-01 2.39549279e-01 3.66371095e-01 3.53266746e-01
5.44737220e-01 6.51817739e-01 3.21677625e-01 -4.41761374e-01
-1.58762050e+00 -1.10553086e+00 -4.10406202e-01 9.83166814e-01
-4.54545349e-01 -1.13567591e+00 -5.33102751e-01 -7.60357201e-01
-1.26895281e-02 7.39639997e-01 -6.59412622e-01 2.00835690e-01
-1.19259691e+00 -6.20533943e-01 9.85237300e-01 4.07401770e-01
2.16250509e-01 -1.09555113e+00 -7.63490200e-01 3.47365826e-01
-4.15101558e-01 -1.04820335e+00 -2.18496248e-01 5.00059843e-01
-6.68798327e-01 -1.37052119e+00 -3.10359091e-01 -7.36035764e-01
2.67653167e-01 -5.55214107e-01 1.35046530e+00 2.17614725e-01
-2.28710309e-01 9.60573331e-02 -5.25487721e-01 -2.91781008e-01
-8.92072141e-01 7.35939205e-01 -7.52827246e-03 -5.46382368e-01
5.73048413e-01 -6.96325243e-01 2.09386513e-01 -1.81549117e-01
-9.87744868e-01 1.77719519e-02 5.19013166e-01 8.42610240e-01
2.59200901e-01 -1.36649400e-01 1.48266092e-01 -1.15043271e+00
6.17925644e-01 -1.65058702e-01 -4.97630864e-01 1.91398889e-01
-6.05875432e-01 2.70157516e-01 6.81664348e-01 -4.72207695e-01
-8.24238896e-01 -7.24928081e-02 -6.94893301e-01 2.86177874e-01
2.70409584e-01 9.96460378e-01 -8.36639628e-02 4.25932020e-01
9.72046077e-01 1.44332930e-01 -2.04636842e-01 -1.04409623e+00
4.53955680e-01 6.50611401e-01 6.12122118e-01 -7.51813173e-01
1.26525986e+00 4.42093104e-01 -5.21837994e-02 -1.04255307e+00
-6.97482169e-01 -4.18397009e-01 -1.04707932e+00 2.11891040e-01
5.22695959e-01 -7.44730949e-01 -4.52397168e-01 -1.06384046e-01
-1.03925848e+00 -3.09083611e-01 -6.93812668e-01 3.55573058e-01
-4.76382434e-01 8.35758448e-01 -7.30846226e-01 -5.11302769e-01
-4.50812936e-01 -6.55242383e-01 1.00732195e+00 -4.48793858e-01
-6.06947422e-01 -1.16112161e+00 1.64332584e-01 6.68553412e-02
9.82355140e-03 1.79306507e-01 1.25622451e+00 -8.93216074e-01
-2.32604086e-01 -1.93924516e-01 9.41875204e-02 2.22652540e-01
2.11463720e-01 -4.16020416e-02 -5.78062713e-01 -1.25820190e-01
-1.50908872e-01 -2.72701830e-01 8.25960875e-01 -2.05004960e-01
2.70704448e-01 -4.17983383e-01 -1.08960509e-01 5.88275671e-01
1.18578565e+00 -3.01251024e-01 6.69563115e-01 8.22632730e-01
4.44422215e-01 3.08281213e-01 4.23231691e-01 5.46278298e-01
1.37849808e-01 6.46324277e-01 -1.45986840e-01 5.40232733e-02
-4.11568701e-01 -5.20063698e-01 4.04215127e-01 1.25748169e+00
-9.43473652e-02 -1.05035506e-01 -1.05039620e+00 9.15512979e-01
-1.56639791e+00 -9.49857295e-01 -2.81152636e-01 1.98412919e+00
1.37559295e+00 2.36119434e-01 1.34439602e-01 5.93812943e-01
1.62328601e-01 1.96758479e-01 8.73693824e-02 -2.25116268e-01
-6.23738587e-01 8.66858900e-01 5.61269522e-01 5.30089915e-01
-1.04473066e+00 1.28033221e+00 7.07471657e+00 1.16415024e+00
-9.35379446e-01 1.76062450e-01 -1.15690701e-01 -3.17581207e-01
-6.28511250e-01 3.25548917e-01 -7.47813702e-01 4.71466124e-01
7.70282447e-01 -3.30915242e-01 3.08432877e-01 6.47268534e-01
1.10575132e-01 3.22453231e-02 -1.17248380e+00 9.11906540e-01
9.85687301e-02 -1.46998060e+00 1.33510917e-01 9.48405489e-02
5.49484134e-01 2.16959864e-01 -2.81671762e-01 1.98206231e-01
2.28073955e-01 -8.76552105e-01 1.09596217e+00 9.10572708e-02
7.07342863e-01 -6.40507460e-01 4.03342485e-01 2.72527099e-01
-1.08186221e+00 1.38898626e-01 -4.70392138e-01 -1.73428014e-01
5.27306020e-01 5.42077839e-01 -9.35551167e-01 5.31358004e-01
2.87237078e-01 7.44457304e-01 -8.37832332e-01 7.38602221e-01
-6.51471376e-01 1.01403308e+00 -9.46020126e-01 -9.61972773e-02
3.25851142e-01 -2.44509935e-01 6.16962254e-01 1.57463670e+00
2.07009658e-01 4.01887558e-02 8.50735232e-03 6.57031417e-01
-1.26182944e-01 7.08444715e-01 -2.66417027e-01 -2.66797245e-01
4.58456308e-01 1.03222370e+00 -8.02627981e-01 -3.83592784e-01
-4.29431498e-01 7.79591560e-01 4.65524077e-01 -3.48675884e-02
-6.04499459e-01 -1.81646213e-01 5.45416653e-01 5.82804143e-01
3.18307400e-01 -5.69899142e-01 -6.81146681e-01 -1.30568790e+00
2.37829238e-01 -9.52027678e-01 6.55662954e-01 -2.60226488e-01
-1.27198279e+00 4.68195438e-01 -6.55799806e-02 -1.02825236e+00
-2.35847697e-01 -8.17308307e-01 -1.75078645e-01 5.99829197e-01
-1.26246500e+00 -1.42726910e+00 3.85026395e-01 4.36120808e-01
5.01552284e-01 -2.53513008e-01 7.64164507e-01 7.15982974e-01
-1.84497356e-01 6.77701414e-01 1.95219561e-01 6.70235872e-01
7.36811757e-01 -1.14776897e+00 8.18968594e-01 1.06317949e+00
8.33311141e-01 8.43931556e-01 7.49448121e-01 -9.78376865e-01
-1.29604983e+00 -6.27066016e-01 1.54921961e+00 -6.14031494e-01
1.19851971e+00 -6.07926965e-01 -8.02935362e-01 7.45790720e-01
1.75989509e-01 -6.95810258e-01 7.57468879e-01 6.24422610e-01
-4.53019321e-01 3.23079884e-01 -4.90177304e-01 9.27188396e-01
1.35528135e+00 -8.04891169e-01 -9.08494473e-01 5.00584960e-01
4.71948870e-02 -4.38752353e-01 -1.00402594e+00 9.07678157e-02
4.56442922e-01 -6.49458289e-01 8.73857856e-01 -5.90107203e-01
6.19833589e-01 5.25585674e-02 -7.31125027e-02 -9.91034150e-01
-5.15458398e-02 -6.80524468e-01 2.13444546e-01 1.35728776e+00
7.24046826e-01 -1.96783453e-01 9.92951274e-01 3.04229468e-01
-3.03959668e-01 -3.40898722e-01 -1.31918228e+00 -1.06676066e+00
4.40616101e-01 -8.39200735e-01 3.08461189e-01 1.13044095e+00
4.24510896e-01 6.84415281e-01 -3.59168142e-01 -4.49163824e-01
3.15055549e-01 5.51396422e-02 6.55782461e-01 -1.31504905e+00
-1.81992337e-01 -6.14232361e-01 -3.05025816e-01 -1.07080865e+00
2.17484102e-01 -1.40147507e+00 -8.49188566e-02 -1.44591177e+00
-7.40778968e-02 -6.70179427e-01 3.94531190e-01 7.57427990e-01
6.89365789e-02 7.53998995e-01 1.94340587e-01 3.31971169e-01
-7.54029900e-02 5.38834691e-01 7.62922823e-01 -1.12517558e-01
-1.53744355e-01 -2.37098917e-01 -5.03495395e-01 6.87809765e-01
5.00867903e-01 -7.19720662e-01 1.91435590e-02 -6.31413400e-01
8.31257522e-01 -5.67341864e-01 -9.27954987e-02 -7.62287676e-01
1.65141150e-01 -1.80869047e-02 -1.62997153e-02 -7.53783345e-01
2.42312655e-01 -7.80096471e-01 2.25141600e-01 4.61590946e-01
-2.48655722e-01 1.65613860e-01 1.08580396e-01 -3.44488882e-02
-3.01361494e-02 -5.34845889e-01 6.25087798e-01 -1.95013404e-01
-4.35820013e-01 1.11192524e-01 -7.27010369e-01 4.09499079e-01
3.23163539e-01 -2.31423765e-01 -3.88276242e-02 -5.00115976e-02
-7.20567405e-01 -2.56340981e-01 5.24055898e-01 2.00095579e-01
9.52958614e-02 -1.38560009e+00 -1.04612422e+00 1.42119855e-01
7.98433945e-02 -1.34895816e-01 -3.01471382e-01 7.14164674e-01
-7.19778597e-01 2.42242306e-01 -5.62147014e-02 -1.19947359e-01
-1.52121174e+00 4.89609331e-01 -6.55525550e-02 -5.02348959e-01
-1.11421490e+00 5.29209316e-01 -4.72741336e-01 -3.85498345e-01
1.14579685e-01 -3.94392401e-01 -4.68027918e-03 5.87962449e-01
1.30352065e-01 2.43008167e-01 4.99985546e-01 -8.08505237e-01
-3.46182138e-01 4.61783141e-01 -9.43815112e-02 -3.18474561e-01
1.61392188e+00 3.35516296e-02 -1.17910653e-01 3.14028919e-01
1.03260303e+00 5.59558332e-01 -5.57031274e-01 -5.04842043e-01
4.66845602e-01 -3.17095935e-01 -1.54156655e-01 -7.36612976e-01
-6.85854316e-01 7.83394992e-01 -1.27026066e-01 1.35412961e-01
8.12784135e-01 4.18698862e-02 9.99082625e-01 7.33507872e-01
6.97401240e-02 -1.77257633e+00 -2.26955876e-01 8.74002695e-01
7.75163054e-01 -6.44698679e-01 5.88123500e-01 -3.95400017e-01
-4.29100841e-01 1.24826574e+00 -1.15457863e-01 -2.08319798e-01
5.65004349e-01 7.76038170e-01 1.65394008e-01 -3.00495207e-01
-5.02455056e-01 -5.30841470e-01 3.07037085e-01 5.36412179e-01
5.91448486e-01 -2.52870888e-01 -1.04533005e+00 6.71699643e-01
-1.03831446e+00 -3.99998605e-01 3.96006405e-01 9.85836267e-01
-1.84910178e-01 -1.56860578e+00 -2.18446255e-01 3.90897483e-01
-5.27069271e-01 -7.92247653e-01 -7.80408204e-01 1.14388287e+00
2.37666219e-01 4.98725533e-01 1.39870197e-01 -2.42860720e-01
4.27677214e-01 3.02168846e-01 6.62089109e-01 -7.75191426e-01
-1.17733240e+00 1.94374070e-01 6.54924512e-01 -4.22483012e-02
-3.23936701e-01 -9.26098824e-01 -1.21201742e+00 -6.56035960e-01
-4.12799329e-01 2.04857171e-01 5.71146905e-01 1.32311010e+00
-2.11143628e-01 2.09431484e-01 7.21977130e-02 -5.33680916e-01
-5.04527688e-01 -1.13460493e+00 -5.48200548e-01 6.45160139e-01
-4.01198208e-01 -3.75665694e-01 -2.46358588e-01 2.03468680e-01]
|
[10.654107093811035, 9.987006187438965]
|
1f8bd7a7-038f-4e32-aeec-02a2f1be9c66
|
bilingual-word-embeddings-from-parallel-and
| null | null |
https://aclanthology.org/N16-1083
|
https://aclanthology.org/N16-1083.pdf
|
Bilingual Word Embeddings from Parallel and Non-parallel Corpora for Cross-Language Text Classification
| null |
['Achim Rettinger', 'Aditya Mogadala']
|
2016-06-01
| null | null | null |
naacl-2016-6
|
['multilingual-word-embeddings']
|
['methodology']
|
[-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01]
|
[-7.3765082359313965, 3.5638673305511475]
|
69855cf6-1a4c-407b-a1b4-155ac53012fc
|
bootstrapping-apprenticeship-learning
| null | null |
http://papers.nips.cc/paper/4160-bootstrapping-apprenticeship-learning
|
http://papers.nips.cc/paper/4160-bootstrapping-apprenticeship-learning.pdf
|
Bootstrapping Apprenticeship Learning
|
We consider the problem of apprenticeship learning where the examples, demonstrated by an expert, cover only a small part of a large state space. Inverse Reinforcement Learning (IRL) provides an efficient tool for generalizing the demonstration, based on the assumption that the expert is maximizing a utility function that is a linear combination of state-action features. Most IRL algorithms use a simple Monte Carlo estimation to approximate the expected feature counts under the expert's policy. In this paper, we show that the quality of the learned policies is highly sensitive to the error in estimating the feature counts. To reduce this error, we introduce a novel approach for bootstrapping the demonstration by assuming that: (i), the expert is (near-)optimal, and (ii), the dynamics of the system is known. Empirical results on gridworlds and car racing problems show that our approach is able to learn good policies from a small number of demonstrations.
|
['Brahim Chaib-Draa', 'Abdeslam Boularias']
|
2010-12-01
| null | null | null |
neurips-2010-12
|
['carracing-v0']
|
['playing-games']
|
[-2.40633860e-01 1.19721323e-01 -3.45187277e-01 6.93408586e-03
-7.74481297e-01 -7.14547098e-01 5.82999945e-01 3.36790197e-02
-7.59876668e-01 1.31705368e+00 -4.28664386e-01 -3.83863777e-01
-2.32257128e-01 -6.12192273e-01 -1.22015989e+00 -6.30414844e-01
-5.31352162e-01 6.02287650e-01 3.30713212e-01 -1.47792384e-01
2.78097242e-01 4.29276407e-01 -1.56446755e+00 -3.72400820e-01
1.01088250e+00 7.06012785e-01 3.70289117e-01 1.10792804e+00
2.82189012e-01 8.35478663e-01 -7.26867735e-01 1.95401147e-01
3.74387145e-01 -6.13996625e-01 -5.45943141e-01 1.64131999e-01
-1.45642415e-01 -6.50145531e-01 -3.44587177e-01 1.28110600e+00
1.36066586e-01 5.02387702e-01 5.83105206e-01 -1.57084811e+00
1.36339560e-01 3.37156564e-01 -1.28681928e-01 -3.32150748e-03
3.95144969e-01 6.94629312e-01 7.72292376e-01 -1.38176337e-01
5.48395216e-01 1.07167292e+00 4.35668081e-01 4.60880488e-01
-1.30563021e+00 -5.96281946e-01 3.13702226e-01 1.41418532e-01
-1.21604002e+00 -6.04554340e-02 3.31484854e-01 -4.09332514e-01
7.73337305e-01 -1.47914216e-01 9.44722891e-01 8.96116436e-01
2.37136841e-01 5.55830419e-01 1.47193074e+00 -4.74738449e-01
7.93070078e-01 4.21495765e-01 -2.91795880e-01 8.04492652e-01
3.08068603e-01 8.75757515e-01 -1.65685326e-01 -4.26100463e-01
7.71338463e-01 -6.66655004e-02 -2.47523472e-01 -1.00568330e+00
-1.01092613e+00 8.85041773e-01 1.61762210e-03 -1.29134387e-01
-7.10716307e-01 4.49978530e-01 1.37210429e-01 8.13911438e-01
-2.40541518e-01 6.57359838e-01 -4.09316808e-01 -5.19873083e-01
-4.61077332e-01 5.73248327e-01 1.21948755e+00 9.36257124e-01
7.29769051e-01 1.38336658e-01 1.47178233e-01 2.40117371e-01
9.97071806e-03 7.02055871e-01 4.07299131e-01 -1.23147893e+00
4.64879394e-01 1.19712964e-01 1.14218044e+00 -6.33044019e-02
5.05093038e-02 -1.06513605e-01 7.70392269e-02 1.04269266e+00
6.39306009e-01 -6.20136619e-01 -7.14664757e-01 1.95250547e+00
3.98382217e-01 3.08453709e-01 3.39160562e-01 5.48856258e-01
-3.67876142e-01 6.58022046e-01 -1.21129379e-01 -3.37076336e-01
6.72973633e-01 -6.18214607e-01 -5.28184891e-01 -1.74851254e-01
4.29199070e-01 -3.04078519e-01 1.07857835e+00 4.05792505e-01
-1.07119572e+00 -4.84938085e-01 -1.21681285e+00 7.61767507e-01
-5.86103648e-02 2.40333714e-02 1.58642814e-01 5.02430141e-01
-7.69345641e-01 1.17831349e+00 -1.01411283e+00 -1.27798796e-01
1.73124328e-01 4.63049412e-01 -1.64202601e-01 8.03957433e-02
-9.34906781e-01 1.13496268e+00 5.31489611e-01 -2.57974684e-01
-1.65871739e+00 -3.08789790e-01 -5.58728695e-01 1.82456613e-01
7.26762772e-01 -3.54024708e-01 1.66042590e+00 -8.53924930e-01
-1.94109130e+00 3.89198400e-02 3.45646769e-01 -5.54073930e-01
8.78860176e-01 -2.40407810e-01 9.17427987e-02 2.50327408e-01
5.44912852e-02 3.18203866e-01 9.92361724e-01 -1.21544027e+00
-9.04722869e-01 -1.30087003e-01 3.62213612e-01 4.33966637e-01
6.44163862e-02 -6.20433331e-01 1.98347524e-01 -1.18635306e-02
-5.10339081e-01 -1.23356080e+00 -4.98840570e-01 7.40901902e-02
-1.17656887e-01 -2.74792850e-01 5.80862939e-01 -2.24923834e-01
6.49260938e-01 -2.16475964e+00 7.63659999e-02 4.10115182e-01
-3.71879190e-01 1.89620972e-01 -2.09473185e-02 6.61314905e-01
6.86981529e-02 -2.25757092e-01 -4.12623324e-02 6.95246784e-03
2.17948407e-01 3.98801118e-01 -4.53507334e-01 6.02521241e-01
-1.65586293e-01 5.15172899e-01 -1.17469203e+00 -2.87105024e-01
3.00501764e-01 -1.15329042e-01 -4.70658302e-01 6.43580556e-01
-4.80451286e-01 4.31927621e-01 -7.99036384e-01 -1.80210859e-01
1.01319440e-01 -2.91071422e-02 4.12787527e-01 4.17050540e-01
-1.84809655e-01 5.82305863e-02 -1.45577621e+00 1.30351615e+00
-6.20390236e-01 2.61539906e-01 4.58055660e-02 -1.04438519e+00
5.61636627e-01 3.09498668e-01 3.73093992e-01 -1.70695469e-01
8.56384188e-02 2.18842670e-01 2.47843400e-01 -7.07645357e-01
-1.15951151e-02 -3.74708772e-01 -6.04310483e-02 7.28726745e-01
2.48302624e-01 -6.69989824e-01 2.85093367e-01 -2.55663060e-02
9.69403863e-01 4.94418591e-01 6.94772303e-01 -1.30892739e-01
2.19322279e-01 2.44009405e-01 3.81411970e-01 1.07889056e+00
-2.88537383e-01 -2.47796759e-01 7.42846787e-01 -1.18030921e-01
-1.16824937e+00 -1.08579469e+00 3.96172702e-01 4.80194956e-01
2.59183794e-01 -1.03959411e-01 -8.76369357e-01 -8.61237526e-01
1.62805736e-01 1.19534516e+00 -6.31078422e-01 -1.95596680e-01
-4.21933860e-01 -1.34861350e-01 -4.21285219e-02 5.03244936e-01
3.67492706e-01 -8.74006152e-01 -1.13977575e+00 2.55909950e-01
1.82583407e-01 -8.60108852e-01 -4.81780857e-01 2.84740776e-01
-9.10412967e-01 -1.34637535e+00 -4.46364015e-01 -4.62594718e-01
7.10852146e-01 -3.39314118e-02 6.97112083e-01 -1.14762172e-01
-1.29535779e-01 8.31741989e-01 8.91222209e-02 -3.74125928e-01
-7.09389150e-01 -3.24615866e-01 2.53415525e-01 -4.50941354e-01
-1.53531268e-01 -5.43331921e-01 -3.36033583e-01 1.35882065e-01
-3.09461176e-01 -2.93327361e-01 5.64719975e-01 1.05086386e+00
3.80874157e-01 3.71187657e-01 4.18987632e-01 -6.07579768e-01
8.97196770e-01 -2.28799492e-01 -1.33876681e+00 2.70378441e-01
-7.15841413e-01 4.72247869e-01 9.91688311e-01 -9.73435640e-01
-9.56131876e-01 3.02129239e-01 2.74551928e-01 -5.07666349e-01
-3.35215807e-01 7.16345906e-02 5.81917502e-02 8.00426025e-03
7.20283568e-01 3.25869530e-01 4.93488580e-01 -2.54557550e-01
1.88776985e-01 3.22069347e-01 2.76087403e-01 -8.98975968e-01
8.43416572e-01 1.23334296e-01 1.43001005e-01 -5.49945772e-01
-4.58276480e-01 -6.47824854e-02 -3.01467448e-01 -3.23454201e-01
4.02651131e-01 -7.79973269e-01 -1.36571395e+00 1.20292135e-01
-6.10868454e-01 -9.34223473e-01 -7.90258348e-01 8.91752660e-01
-1.31103826e+00 2.21926361e-01 -2.53035367e-01 -1.42303455e+00
3.76335621e-01 -1.12581193e+00 4.68572229e-01 4.24821436e-01
6.87807798e-02 -7.57551372e-01 3.13682050e-01 -3.19700122e-01
1.54417664e-01 1.87850133e-01 8.98188114e-01 -6.60945475e-01
-6.48317099e-01 -2.64714003e-01 3.88908654e-01 3.69486243e-01
-7.69265145e-02 -2.11579800e-01 -5.56271493e-01 -5.53083360e-01
9.77361128e-02 -6.32533252e-01 2.45127484e-01 3.07419449e-01
9.28769112e-01 -6.56888723e-01 -3.73468101e-01 -1.12717144e-01
1.37880254e+00 6.67550147e-01 2.45212480e-01 3.60930860e-01
-4.72187363e-02 4.17420626e-01 1.14516103e+00 6.69172287e-01
1.37747258e-01 4.78717774e-01 5.26299357e-01 3.86172503e-01
4.66253221e-01 -6.70251012e-01 5.22833109e-01 -6.18025437e-02
1.74960066e-02 1.20318696e-01 -4.39251155e-01 4.93037969e-01
-2.08353281e+00 -1.01409411e+00 5.71721017e-01 2.68313122e+00
8.36326122e-01 2.57835418e-01 5.22613466e-01 -1.62893772e-01
6.62209272e-01 -3.97849321e-01 -1.13088393e+00 -2.34222531e-01
6.17809355e-01 1.06798016e-01 7.19725490e-01 6.99951828e-01
-6.96626306e-01 4.97770339e-01 7.04686451e+00 5.86704552e-01
-6.75679445e-01 -1.50131732e-01 2.11666092e-01 -1.11538852e-02
2.22920895e-01 2.06460819e-01 -7.41300464e-01 7.15246677e-01
9.54781115e-01 -5.55838645e-01 8.80749941e-01 1.33388734e+00
2.26572648e-01 -6.00884616e-01 -1.38684082e+00 5.07129371e-01
-4.27949429e-01 -8.74398947e-01 -5.93768060e-01 1.85744807e-01
7.11680830e-01 -9.08408090e-02 1.30981095e-02 7.75440991e-01
1.03124559e+00 -6.60733342e-01 6.88259661e-01 4.16689098e-01
6.24701917e-01 -1.02307570e+00 6.26603901e-01 1.01499212e+00
-7.17735767e-01 -5.26689410e-01 -1.35446697e-01 -1.51238635e-01
4.30892967e-02 -1.47467956e-01 -1.04393196e+00 1.27682000e-01
3.14334989e-01 1.86206147e-01 9.34606940e-02 1.12687004e+00
-4.83047605e-01 4.22907531e-01 -5.65531552e-01 -6.08891070e-01
3.80294800e-01 -3.67984682e-01 4.90157306e-01 6.58679605e-01
4.07345712e-01 1.11119762e-01 6.80160224e-01 7.91268587e-01
4.35738057e-01 -2.77799040e-01 -8.34809899e-01 -7.72724375e-02
3.93089682e-01 8.16418231e-01 -5.20216882e-01 -4.90169287e-01
-8.28028396e-02 6.07731044e-01 4.71365005e-01 4.90638942e-01
-8.56355548e-01 -4.64398742e-01 6.56321108e-01 -2.89443940e-01
6.11465394e-01 -2.85394639e-01 4.60474700e-01 -8.54282200e-01
-1.48791254e-01 -7.99178004e-01 2.37478763e-01 -5.85511565e-01
-8.32357824e-01 1.40883952e-01 3.18226486e-01 -1.30175710e+00
-9.19288337e-01 -5.01914680e-01 -5.68841696e-01 6.92037344e-01
-1.34678602e+00 -2.23610744e-01 1.00642502e-01 5.03197372e-01
4.59777087e-01 -9.22886059e-02 7.09294319e-01 -4.49832559e-01
-3.00683826e-01 1.43645361e-01 3.24436694e-01 -2.18132272e-01
1.77925780e-01 -1.49011981e+00 3.21851671e-02 5.61884403e-01
-6.69007450e-02 2.86253065e-01 1.28697932e+00 -4.91695404e-01
-1.32476354e+00 -6.76523745e-01 7.78369531e-02 -6.55663162e-02
8.67287219e-01 -1.32454475e-02 -6.40957713e-01 7.93547153e-01
-4.88059558e-02 8.62026811e-02 9.31208134e-02 -1.41533658e-01
1.08540773e-01 -4.90975082e-02 -1.39277041e+00 6.23026550e-01
6.13301098e-01 -2.60085702e-01 -6.48237586e-01 4.22396123e-01
2.83425152e-01 -4.92853850e-01 -6.95854843e-01 1.17269926e-01
5.58413804e-01 -7.53895104e-01 6.99808002e-01 -9.21122193e-01
2.19228286e-02 -2.42119327e-01 1.16239637e-01 -1.92875814e+00
6.27341047e-02 -8.40617061e-01 -4.52082485e-01 4.51408416e-01
5.10671027e-02 -8.22402179e-01 7.02945054e-01 4.47430253e-01
3.16338927e-01 -6.41915798e-01 -1.13003659e+00 -1.28657746e+00
5.30092977e-02 9.01311710e-02 3.33407223e-01 3.03329885e-01
3.50802392e-01 1.91093400e-01 -3.99948031e-01 2.15555295e-01
8.41966629e-01 2.92554438e-01 9.52507317e-01 -9.46509302e-01
-8.13933909e-01 -1.04098849e-01 -7.33833909e-02 -1.21167040e+00
5.31277835e-01 -1.45884424e-01 5.48421860e-01 -1.17494512e+00
1.47752628e-01 -4.92430001e-01 -1.46264896e-01 2.02970132e-01
-4.59360145e-02 -6.68389559e-01 1.48547336e-01 -1.10910036e-01
-7.12309122e-01 6.35909259e-01 1.39213514e+00 2.31497064e-01
-2.52169847e-01 6.77549720e-01 -2.03460440e-01 6.97052479e-01
9.21481907e-01 -5.50724685e-01 -7.28321612e-01 1.60786822e-01
-3.31352316e-02 7.18450487e-01 3.09892714e-01 -9.33141232e-01
1.90130368e-01 -6.01904273e-01 1.21596396e-01 -3.69167775e-01
5.21374047e-01 -9.57430005e-01 9.08698328e-03 1.11108804e+00
-5.36009252e-01 1.46234464e-02 6.77232146e-02 1.07436049e+00
1.15389712e-01 -6.42379701e-01 8.96360397e-01 -3.19705158e-01
-4.97376680e-01 7.22989812e-02 -6.90039873e-01 1.20174326e-01
1.26316309e+00 7.86694139e-02 -2.56887428e-03 -8.17483068e-01
-7.44196475e-01 5.92519701e-01 5.42910933e-01 -4.95850518e-02
4.03085381e-01 -9.15735126e-01 -2.67384410e-01 2.50128261e-03
-1.32930711e-01 -4.09013778e-01 -7.67440572e-02 5.33533573e-01
3.27487290e-02 1.94712341e-01 -2.56292552e-01 -4.11464334e-01
-1.09242606e+00 7.42905736e-01 5.50060630e-01 -2.91008562e-01
-6.27990603e-01 2.67655075e-01 -1.30256876e-01 -1.87553242e-01
3.52535903e-01 -2.57999986e-01 5.62472492e-02 -5.06538689e-01
4.10970300e-01 4.31198388e-01 -4.62290257e-01 1.14631606e-02
1.08115070e-01 2.70132840e-01 1.51348114e-02 -5.80445528e-01
1.29391050e+00 4.72407974e-02 6.15831971e-01 6.28171802e-01
7.55109489e-01 -4.13905114e-01 -2.11188221e+00 -8.42436180e-02
-3.88597362e-02 -6.16085410e-01 -1.95944160e-01 -5.97564161e-01
-3.86452675e-01 6.68500602e-01 6.79418981e-01 2.05805302e-01
6.85530543e-01 -3.39528024e-01 1.73399165e-01 8.32351029e-01
8.57095003e-01 -1.41847122e+00 2.97922701e-01 2.09481895e-01
5.44859469e-01 -1.04311097e+00 1.27552569e-01 1.63026482e-01
-8.37053895e-01 1.11882031e+00 6.73726976e-01 -6.82989299e-01
4.60226774e-01 3.79888952e-01 -3.97107482e-01 1.84719965e-01
-9.88929570e-01 -1.97556198e-01 -2.46276587e-01 6.96291566e-01
-3.64859283e-01 2.26647347e-01 1.18009262e-02 4.38345186e-02
-2.44799796e-02 1.98344067e-01 7.50523508e-01 1.11406553e+00
-9.42033589e-01 -8.73585939e-01 -3.71704876e-01 3.34424824e-01
-1.30570948e-01 5.12508035e-01 8.85559618e-02 1.16714859e+00
-2.84389377e-01 7.58832395e-01 -1.39833748e-01 2.05694184e-01
4.03134704e-01 3.58089320e-02 9.02362466e-01 -5.58289766e-01
-9.01832059e-02 -6.33777454e-02 6.19785078e-02 -7.57786095e-01
-8.42593312e-02 -7.31268287e-01 -1.12744749e+00 -3.43665741e-02
-3.49663556e-01 5.99948704e-01 7.66608417e-01 9.95606244e-01
-1.16969552e-02 3.37386370e-01 9.82111216e-01 -7.20427811e-01
-1.68878329e+00 -7.58059323e-01 -8.18016291e-01 1.26723096e-01
6.19545281e-01 -1.03655183e+00 -6.70106709e-01 -3.76917183e-01]
|
[4.257894515991211, 2.0640571117401123]
|
4eb19b71-07c4-42e8-8a27-54a25454eeaa
|
structure-inference-machines-recurrent-neural
|
1511.04196
| null |
http://arxiv.org/abs/1511.04196v2
|
http://arxiv.org/pdf/1511.04196v2.pdf
|
Structure Inference Machines: Recurrent Neural Networks for Analyzing Relations in Group Activity Recognition
|
Rich semantic relations are important in a variety of visual recognition
problems. As a concrete example, group activity recognition involves the
interactions and relative spatial relations of a set of people in a scene.
State of the art recognition methods center on deep learning approaches for
training highly effective, complex classifiers for interpreting images.
However, bridging the relatively low-level concepts output by these methods to
interpret higher-level compositional scenes remains a challenge. Graphical
models are a standard tool for this task. In this paper, we propose a method to
integrate graphical models and deep neural networks into a joint framework.
Instead of using a traditional inference method, we use a sequential inference
modeled by a recurrent neural network. Beyond this, the appropriate structure
for inference can be learned by imposing gates on edges between nodes.
Empirical results on group activity recognition demonstrate the potential of
this model to handle highly structured learning tasks.
|
['Zhiwei Deng', 'Greg Mori', 'Arash Vahdat', 'Hexiang Hu']
|
2015-11-13
|
structure-inference-machines-recurrent-neural-1
|
http://openaccess.thecvf.com/content_cvpr_2016/html/Deng_Structure_Inference_Machines_CVPR_2016_paper.html
|
http://openaccess.thecvf.com/content_cvpr_2016/papers/Deng_Structure_Inference_Machines_CVPR_2016_paper.pdf
|
cvpr-2016-6
|
['group-activity-recognition']
|
['computer-vision']
|
[ 3.98429692e-01 -1.03394119e-02 -1.73947036e-01 -7.02646613e-01
-7.68207908e-02 -3.99880677e-01 1.06286156e+00 1.52537838e-01
-1.71040580e-01 5.06223202e-01 3.37266356e-01 -4.36106712e-01
-1.78737789e-01 -9.06102479e-01 -6.07258081e-01 -5.50466359e-01
5.66792451e-02 3.91455084e-01 7.72764683e-02 1.69925429e-02
2.79145420e-01 7.28380680e-01 -1.54650176e+00 6.78517640e-01
5.40392101e-01 8.00555885e-01 -7.90800974e-02 5.53837657e-01
-3.57840866e-01 1.34285188e+00 -7.21716940e-01 -3.71683151e-01
-2.22166479e-01 -8.57945383e-01 -8.27322066e-01 3.91115576e-01
6.43252313e-01 -2.03155994e-01 -3.24750602e-01 8.71462226e-01
-2.43725590e-02 2.67461091e-01 9.02216792e-01 -1.11780572e+00
-5.89842498e-01 6.34918034e-01 -3.01649958e-01 1.98107526e-01
3.68993312e-01 -1.17551528e-01 1.28841043e+00 -8.03585827e-01
4.61929411e-01 1.34274828e+00 5.24256647e-01 3.27814639e-01
-1.62588811e+00 -4.28804070e-01 5.29873490e-01 4.46861267e-01
-1.29210484e+00 -2.89270103e-01 7.68659830e-01 -5.35243928e-01
1.01281238e+00 2.34831378e-01 8.74119222e-01 1.17961633e+00
-1.58160567e-01 7.11563170e-01 1.10129440e+00 -4.37198341e-01
3.46411288e-01 -1.43881217e-01 3.30624700e-01 8.91661167e-01
3.41490000e-01 -2.90894032e-01 -8.15203011e-01 8.00471604e-02
1.14998651e+00 3.53834808e-01 -9.08266380e-02 -3.83484274e-01
-1.08449566e+00 9.18062389e-01 8.37522984e-01 5.71318686e-01
-1.70961618e-01 5.74498296e-01 7.06146508e-02 4.67190370e-02
3.93826455e-01 5.03870487e-01 2.35835060e-01 3.09010416e-01
-8.69975269e-01 1.34684235e-01 1.01351011e+00 4.86425608e-01
7.37573802e-01 -9.53328684e-02 -4.90177602e-01 6.41264915e-01
5.30955255e-01 -2.27241330e-02 2.78831203e-03 -9.56420362e-01
2.32206434e-01 9.21470404e-01 -1.57758102e-01 -1.13709509e+00
-2.93883711e-01 -6.98772132e-01 -8.80489409e-01 3.28779995e-01
5.57227910e-01 2.75594920e-01 -9.73873258e-01 1.77193439e+00
-5.07699363e-02 2.91734755e-01 -1.54897079e-01 8.76207829e-01
6.02966666e-01 6.86202407e-01 1.32215098e-01 1.51354447e-01
1.41355932e+00 -1.11629903e+00 -7.39729404e-01 -4.51013863e-01
5.78750730e-01 -9.41310972e-02 8.45948219e-01 2.86088049e-01
-1.04870391e+00 -4.42395002e-01 -1.04089797e+00 -3.94075334e-01
-6.84499443e-01 1.97074533e-01 9.36278880e-01 3.54548484e-01
-1.12577188e+00 4.03854281e-01 -8.01909864e-01 -5.14564693e-01
8.50092292e-01 3.62454176e-01 -2.78914303e-01 -1.40947536e-01
-1.00312889e+00 9.54425871e-01 1.44901425e-01 3.99894148e-01
-8.36764455e-01 -3.10984552e-01 -9.47686732e-01 3.55372608e-01
3.76949698e-01 -1.03706956e+00 1.16606545e+00 -1.27973485e+00
-1.41372526e+00 9.99516368e-01 -3.71540904e-01 -4.42037642e-01
3.96834105e-01 -2.40364179e-01 1.36591811e-02 2.79378325e-01
-1.29049435e-01 4.65978086e-01 6.88705981e-01 -1.37523723e+00
-3.18983406e-01 -4.25577104e-01 3.12780797e-01 4.59233999e-01
-2.00576127e-01 1.65006503e-01 -3.91070545e-01 -4.46504265e-01
2.11079523e-01 -5.99790573e-01 -3.93015027e-01 3.20234537e-01
-2.93741435e-01 -3.81471187e-01 7.08707750e-01 -4.85489160e-01
1.08421409e+00 -1.95309103e+00 5.11627614e-01 4.74451274e-01
4.64521021e-01 1.47011220e-01 9.72186029e-02 4.37631190e-01
4.84251156e-02 1.45443723e-01 -3.18913460e-01 -5.07807076e-01
1.24037556e-01 4.56924558e-01 -4.63687241e-01 2.85498381e-01
3.96269232e-01 1.10643780e+00 -8.76524448e-01 -1.45524934e-01
2.16186240e-01 6.04316533e-01 -5.60445309e-01 2.54172504e-01
-3.74564558e-01 4.75737691e-01 -2.81196177e-01 2.89631009e-01
6.47098199e-02 -7.26457357e-01 6.65590167e-01 -1.44334242e-01
1.39292657e-01 5.39069772e-01 -9.95234549e-01 1.78073680e+00
-3.71459931e-01 7.87351668e-01 -1.22876547e-01 -1.54882634e+00
7.41059840e-01 6.73807487e-02 -6.64654747e-02 -5.01374662e-01
1.17437758e-01 -9.04872268e-02 1.02053061e-01 -4.29463446e-01
1.16390260e-02 -1.14051335e-01 -1.30500436e-01 6.24479711e-01
5.18709831e-02 8.19943622e-02 2.43790373e-01 2.58677453e-01
1.03056026e+00 1.10807762e-01 4.05634135e-01 -6.43050447e-02
4.32883739e-01 -2.90386349e-01 2.67745912e-01 9.87906992e-01
2.38956437e-01 4.08908635e-01 7.26575375e-01 -5.75623631e-01
-7.60289967e-01 -1.22667348e+00 2.11479053e-01 1.22191930e+00
-8.22331698e-04 -5.54828942e-01 -7.91338086e-01 -4.59616303e-01
-3.21221322e-01 5.41936874e-01 -7.28045523e-01 -1.10928297e-01
-5.18762469e-01 -5.58694124e-01 6.17320597e-01 8.58760178e-01
6.69796288e-01 -9.67995465e-01 -7.17893839e-01 -2.47062165e-02
-7.18103275e-02 -1.25902760e+00 2.19685379e-02 1.63835213e-01
-7.62561023e-01 -9.27777171e-01 -2.72668034e-01 -8.37784886e-01
1.10936415e+00 1.85469270e-01 1.23408031e+00 3.37871820e-01
-4.33648616e-01 2.96265006e-01 -1.37514845e-01 -9.06406939e-02
-1.16713800e-01 5.83193789e-04 -2.35014647e-01 3.87480468e-01
2.83470958e-01 -7.28707790e-01 -4.12686974e-01 1.08251661e-01
-9.10315871e-01 5.54247200e-01 6.66815400e-01 1.04458535e+00
2.33582765e-01 -1.91831023e-01 2.13479444e-01 -9.72728848e-01
6.54875457e-01 -2.84526706e-01 -5.10922790e-01 4.72168505e-01
-6.32965118e-02 3.03930312e-01 4.19366449e-01 -3.72560799e-01
-1.13668740e+00 6.94736093e-02 1.67929679e-01 -2.93196142e-01
-3.78904313e-01 8.20179880e-01 -1.81059942e-01 1.82943821e-01
6.00924313e-01 2.26344869e-01 -8.89280159e-03 -2.36647859e-01
4.48385924e-01 2.47952983e-01 4.88723308e-01 -6.34447575e-01
5.93384504e-01 6.27141774e-01 3.12496305e-01 -8.51679504e-01
-1.14264929e+00 -1.62202910e-01 -8.96852493e-01 -1.96449295e-01
1.21391249e+00 -8.30446541e-01 -8.67639482e-01 3.45742375e-01
-1.35918438e+00 -5.45589864e-01 -8.08399618e-02 4.00654852e-01
-3.52794409e-01 1.31147370e-01 -6.13007128e-01 -7.74763107e-01
2.49546275e-01 -8.64938855e-01 1.01730072e+00 1.03415705e-01
-5.23790658e-01 -1.40772784e+00 -1.44137740e-01 5.98244309e-01
2.77839035e-01 1.98284104e-01 9.28191245e-01 -5.61329782e-01
-1.01416230e+00 -1.05589472e-01 -3.69492024e-01 1.05441742e-01
2.01907411e-01 -8.54919776e-02 -9.83278990e-01 1.76219448e-01
-5.26581518e-02 -3.84377092e-01 1.13648713e+00 9.21931416e-02
1.61152565e+00 -3.47275734e-01 -4.34269071e-01 6.12571180e-01
1.09943366e+00 -2.96984073e-02 7.07945526e-01 -2.31978714e-01
1.02790618e+00 8.14050615e-01 -1.39249396e-02 1.40423372e-01
3.81944805e-01 7.67730355e-01 2.96056509e-01 -2.54836172e-01
-9.52338651e-02 -4.14213389e-01 2.50802249e-01 5.15181243e-01
-3.84433627e-01 -1.30872026e-01 -1.17508745e+00 1.43717527e-01
-2.20952153e+00 -1.37401307e+00 -8.07351843e-02 1.87746382e+00
7.50998914e-01 -1.49881290e-02 -2.28635520e-01 1.02339119e-01
4.76730257e-01 2.16926217e-01 -4.40029263e-01 -3.05090159e-01
-9.59945992e-02 1.99705407e-01 -1.08014174e-01 5.45933366e-01
-9.18249607e-01 9.92773116e-01 6.91857290e+00 4.99920726e-01
-9.32199478e-01 -8.69032890e-02 6.90189600e-01 -4.99494597e-02
-3.05785090e-01 1.26616210e-01 -6.94125354e-01 4.02287655e-02
7.71812260e-01 3.07437420e-01 5.98845303e-01 3.12946409e-01
1.50042102e-01 -2.74985731e-01 -1.55159783e+00 1.18224633e+00
4.01250392e-01 -1.76865101e+00 4.54484522e-01 1.22726321e-01
5.08143067e-01 -5.32986462e-01 1.17336167e-02 1.28141180e-01
5.09965122e-01 -1.56058884e+00 6.43442512e-01 6.82125270e-01
3.28936666e-01 -2.46937677e-01 3.05324495e-01 3.77226263e-01
-1.07626402e+00 -8.45689029e-02 2.63780318e-02 -7.21670985e-01
7.24622756e-02 3.91155869e-01 -6.58779740e-01 3.29185575e-01
5.21771729e-01 8.87136936e-01 -6.59334123e-01 7.07051516e-01
-7.66239583e-01 4.94253606e-01 -1.70982927e-01 -1.98458746e-01
1.79521799e-01 -3.91058564e-01 6.13210239e-02 1.31083930e+00
1.56432521e-02 1.06070682e-01 7.75647610e-02 1.33572721e+00
-2.03194097e-01 -2.96295553e-01 -1.00782847e+00 -2.14198455e-01
1.61147267e-01 1.19116139e+00 -9.91108775e-01 -5.89334071e-01
-5.52865088e-01 1.11812079e+00 8.13965917e-01 6.70037806e-01
-7.32679546e-01 2.48287991e-02 6.03732049e-01 1.27054811e-01
2.35580951e-01 -5.77932775e-01 -4.71074015e-01 -1.29648387e+00
-8.28871503e-02 -7.58507669e-01 3.78665358e-01 -9.37108636e-01
-1.22830915e+00 1.77001640e-01 1.30467147e-01 -8.10007811e-01
-3.65505427e-01 -8.68809998e-01 -7.66884625e-01 7.24417865e-01
-1.17860925e+00 -1.30473113e+00 -4.83365953e-01 6.29927039e-01
4.59555715e-01 1.59038022e-01 8.85316014e-01 -2.16514990e-02
-5.72618186e-01 1.36553749e-01 -3.78087878e-01 5.14837325e-01
9.49834213e-02 -1.31592762e+00 2.89089262e-01 9.10152733e-01
7.36894727e-01 9.45389867e-01 5.15292883e-01 -3.18152070e-01
-1.22490919e+00 -8.12875569e-01 9.66695487e-01 -5.23449779e-01
6.40560091e-01 -1.01711798e+00 -1.01711321e+00 9.90582526e-01
3.45937937e-01 1.15491502e-01 9.54610527e-01 4.18250591e-01
-8.07564020e-01 2.28539482e-03 -5.05538523e-01 9.54437077e-01
1.45713115e+00 -1.04986310e+00 -7.74917603e-01 1.73622534e-01
2.47963786e-01 -1.60280988e-01 -4.70736265e-01 1.61913946e-01
4.89226937e-01 -9.84074473e-01 9.86117125e-01 -7.89491475e-01
5.17389297e-01 -5.46212256e-01 1.59437023e-02 -1.22106254e+00
-3.61399472e-01 -1.38851598e-01 -1.82000920e-01 8.59407961e-01
3.61260980e-01 -5.88396728e-01 6.94833338e-01 6.14478111e-01
5.19095547e-02 -5.13401687e-01 -4.57120866e-01 -3.70894760e-01
-3.55771065e-01 -2.89078772e-01 3.22889090e-02 1.06936038e+00
1.63710073e-01 9.35987473e-01 -2.48694941e-01 1.61792651e-01
5.30573308e-01 2.54537404e-01 7.25808322e-01 -1.35929406e+00
-6.28711760e-01 -7.18606412e-01 -5.46072602e-01 -1.12090170e+00
5.75677991e-01 -1.06836307e+00 -4.49471250e-02 -2.03030610e+00
3.64828944e-01 3.52897160e-02 -6.86693713e-02 6.64405286e-01
-1.55012542e-02 2.44526610e-01 4.14534509e-02 8.28568339e-02
-7.33874857e-01 4.08268362e-01 9.86472726e-01 -2.71679044e-01
1.00205421e-01 -1.86636984e-01 -6.30210102e-01 9.06625688e-01
7.18818724e-01 -5.80826066e-02 -6.80600226e-01 -6.35253310e-01
5.67311406e-01 -1.75269336e-01 8.48624647e-01 -8.73389959e-01
6.34013176e-01 -1.82561174e-01 6.18276477e-01 -1.44740731e-01
5.64243197e-01 -6.99810863e-01 2.44155467e-01 4.57253635e-01
-6.09761834e-01 -1.97261795e-01 -2.11490802e-02 6.14218950e-01
-5.33951521e-01 4.61319275e-02 3.75077009e-01 -2.91983932e-01
-6.93933964e-01 1.67339891e-01 -6.04095042e-01 -3.18386197e-01
8.45958292e-01 -4.36082453e-01 -4.01326150e-01 -5.59142649e-01
-1.09858418e+00 3.72959264e-02 1.89764991e-01 2.85910100e-01
7.48063684e-01 -1.23013604e+00 -3.26576501e-01 2.60584742e-01
7.59096667e-02 1.01188570e-01 -6.09938093e-02 8.22479427e-01
-5.39891720e-01 2.61442274e-01 -3.12007457e-01 -6.25463367e-01
-1.25152218e+00 2.07896262e-01 6.76360130e-01 -3.06319416e-01
-3.49893123e-01 7.06441224e-01 6.62871242e-01 -1.11611813e-01
2.80782163e-01 -4.72217947e-01 -2.42943496e-01 2.03368068e-01
5.85481703e-01 -5.24282120e-02 -2.15926185e-01 -5.36606431e-01
-3.54168057e-01 4.52265918e-01 8.20178539e-02 -3.02344233e-01
1.37139702e+00 1.27826631e-01 -4.10632163e-01 8.64692628e-01
9.02363479e-01 -3.60886335e-01 -1.29290617e+00 -2.73250639e-01
2.50236958e-01 -1.27935320e-01 -1.32889196e-01 -5.94066143e-01
-7.94904590e-01 1.33979714e+00 6.11984320e-02 2.57968366e-01
9.49557602e-01 2.72633255e-01 -1.79558545e-01 6.51355088e-01
7.68857971e-02 -8.59490335e-01 5.75566411e-01 5.79711020e-01
9.31169331e-01 -1.09239447e+00 -3.05720847e-02 -6.31446421e-01
-3.19983780e-01 1.15336227e+00 5.65526903e-01 -1.88933969e-01
4.87981856e-01 1.45166874e-01 -2.72463799e-01 -4.67175931e-01
-9.06318247e-01 -3.37596148e-01 3.05788130e-01 3.77006173e-01
6.08718634e-01 -4.25445549e-02 -7.49169961e-02 3.01623702e-01
2.47897491e-01 -3.51510420e-02 2.01716453e-01 9.22327399e-01
-2.97714442e-01 -9.53724027e-01 -1.76657736e-01 3.88406426e-01
-2.45654449e-01 -1.32716581e-01 -7.06539810e-01 4.51005995e-01
-8.69908258e-02 7.15638638e-01 4.56694514e-01 -1.24075487e-01
1.55206653e-03 2.10102215e-01 7.94221103e-01 -9.15683687e-01
-5.70599794e-01 -1.85584933e-01 2.48510733e-01 -6.01465583e-01
-8.09650004e-01 -6.04929209e-01 -1.19916999e+00 2.02774610e-02
2.17017367e-01 -1.88152626e-01 4.63993996e-01 1.24744546e+00
1.37442946e-01 6.67291641e-01 1.52841240e-01 -8.68313491e-01
-2.32221335e-01 -7.30662644e-01 -4.75115806e-01 7.40493774e-01
1.94273144e-01 -6.08247280e-01 -2.18849361e-01 3.30200315e-01]
|
[10.01307487487793, 1.127840518951416]
|
c0e834fe-6415-4e70-b52a-c689614a74dc
|
molo-motion-augmented-long-short-contrastive
|
2304.00946
| null |
https://arxiv.org/abs/2304.00946v1
|
https://arxiv.org/pdf/2304.00946v1.pdf
|
MoLo: Motion-augmented Long-short Contrastive Learning for Few-shot Action Recognition
|
Current state-of-the-art approaches for few-shot action recognition achieve promising performance by conducting frame-level matching on learned visual features. However, they generally suffer from two limitations: i) the matching procedure between local frames tends to be inaccurate due to the lack of guidance to force long-range temporal perception; ii) explicit motion learning is usually ignored, leading to partial information loss. To address these issues, we develop a Motion-augmented Long-short Contrastive Learning (MoLo) method that contains two crucial components, including a long-short contrastive objective and a motion autodecoder. Specifically, the long-short contrastive objective is to endow local frame features with long-form temporal awareness by maximizing their agreement with the global token of videos belonging to the same class. The motion autodecoder is a lightweight architecture to reconstruct pixel motions from the differential features, which explicitly embeds the network with motion dynamics. By this means, MoLo can simultaneously learn long-range temporal context and motion cues for comprehensive few-shot matching. To demonstrate the effectiveness, we evaluate MoLo on five standard benchmarks, and the results show that MoLo favorably outperforms recent advanced methods. The source code is available at https://github.com/alibaba-mmai-research/MoLo.
|
['Nong Sang', 'Deli Zhao', 'Yingya Zhang', 'Changxin Gao', 'Zhiwu Qing', 'Shiwei Zhang', 'Xiang Wang']
|
2023-04-03
| null |
http://openaccess.thecvf.com//content/CVPR2023/html/Wang_MoLo_Motion-Augmented_Long-Short_Contrastive_Learning_for_Few-Shot_Action_Recognition_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Wang_MoLo_Motion-Augmented_Long-Short_Contrastive_Learning_for_Few-Shot_Action_Recognition_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['few-shot-action-recognition', 'action-recognition-in-videos']
|
['computer-vision', 'computer-vision']
|
[-8.75184219e-03 -6.12924695e-01 -6.30130947e-01 -2.83384204e-01
-7.38415122e-01 5.28170392e-02 5.49470544e-01 -4.62296784e-01
-3.16537291e-01 4.52659518e-01 5.81265628e-01 3.45045000e-01
-3.83288227e-02 -4.09406632e-01 -5.44090152e-01 -9.01400387e-01
6.56670406e-02 -1.46030977e-01 4.36585009e-01 -1.82191983e-01
2.58293688e-01 1.56054214e-01 -1.59784055e+00 4.34842139e-01
5.07183433e-01 1.19302011e+00 2.54580259e-01 5.53051174e-01
2.23075122e-01 1.30048013e+00 -1.03404887e-01 7.13424385e-02
3.23270649e-01 -7.25280404e-01 -6.01095915e-01 2.26945758e-01
5.87456465e-01 -8.18026602e-01 -1.07098591e+00 9.02827501e-01
5.12243927e-01 5.46659112e-01 3.63550544e-01 -1.29814291e+00
-5.17729342e-01 8.73879641e-02 -6.00847721e-01 6.77334964e-01
5.37894547e-01 7.74515808e-01 1.10158813e+00 -1.13736498e+00
7.70647287e-01 1.32059252e+00 3.81233990e-01 6.56635821e-01
-9.46534872e-01 -3.66374969e-01 3.92328948e-01 7.20736921e-01
-1.30109417e+00 -7.35114038e-01 8.97242725e-01 -5.54135501e-01
9.40277755e-01 3.45982388e-02 8.89949977e-01 1.19364393e+00
2.85419405e-01 1.07998300e+00 7.31260598e-01 -1.53504522e-03
1.73172787e-01 -5.76025844e-01 3.33984233e-02 8.01565170e-01
-2.39288166e-01 4.86146539e-01 -9.38552558e-01 2.06677482e-01
9.22310770e-01 4.08194393e-01 -4.74299133e-01 -4.51537877e-01
-1.23634112e+00 6.08148515e-01 4.88776892e-01 3.35752547e-01
-3.68960649e-01 4.60431069e-01 4.88068014e-01 2.29525000e-01
3.22447658e-01 -1.45030499e-01 -9.37963650e-02 -4.78222072e-01
-7.72620678e-01 2.69115657e-01 2.64185935e-01 6.59296215e-01
9.16734397e-01 2.09157899e-01 -4.35596287e-01 5.47774315e-01
3.45531672e-01 3.32384586e-01 6.14400089e-01 -1.37813973e+00
5.32207668e-01 3.92049849e-01 4.04676124e-02 -1.32979953e+00
-8.35231394e-02 -1.16031192e-01 -8.03125322e-01 2.79089391e-01
4.35727865e-01 1.35353833e-01 -6.95616663e-01 1.79210138e+00
3.13090712e-01 8.09744477e-01 -1.63468421e-01 1.31926441e+00
8.12596917e-01 7.03990042e-01 4.70408518e-03 -3.24505210e-01
1.02892458e+00 -1.32607210e+00 -8.42317939e-01 -4.44930911e-01
5.69769442e-01 -4.67634231e-01 1.16004395e+00 -5.42005301e-02
-1.15601575e+00 -8.48731756e-01 -9.19458330e-01 -1.76727071e-01
8.21670368e-02 -3.23483646e-02 5.25737703e-01 2.85681486e-02
-7.70919085e-01 6.74989820e-01 -1.10775983e+00 -2.01324090e-01
6.32893145e-01 1.47786448e-02 -2.56379902e-01 -3.72540891e-01
-1.15879476e+00 5.98722339e-01 2.71982580e-01 1.18570179e-01
-1.10407329e+00 -6.96101487e-01 -1.09480214e+00 2.86421701e-02
5.24816513e-01 -8.15326512e-01 1.15309680e+00 -1.03591561e+00
-1.48396504e+00 5.96452832e-01 -2.73005962e-01 -3.40259135e-01
6.18496716e-01 -3.92514467e-01 -3.19625109e-01 4.41724032e-01
1.34419829e-01 7.23297119e-01 8.64216149e-01 -7.49092281e-01
-7.73684561e-01 -2.56492108e-01 1.36850268e-01 3.21296722e-01
-2.84403622e-01 -1.00875929e-01 -6.29104614e-01 -9.34327602e-01
1.13090323e-02 -7.02891529e-01 -9.21420380e-02 4.55092400e-01
1.22589335e-01 -1.18914962e-01 9.25844193e-01 -5.49043596e-01
1.34999549e+00 -2.29616833e+00 2.81615287e-01 -4.21743512e-01
2.54476428e-01 3.95815849e-01 -2.14568064e-01 2.43940219e-01
7.05195963e-02 -5.43337703e-01 -1.02608182e-01 -3.28606635e-01
-1.53592408e-01 1.11251250e-01 -3.66668135e-01 6.69297338e-01
1.27578422e-01 1.17345273e+00 -1.19497919e+00 -5.70610762e-01
6.40371561e-01 4.69118357e-01 -5.67992628e-01 3.88815105e-01
-8.44956413e-02 6.14818156e-01 -4.93531168e-01 7.80866086e-01
2.64985293e-01 -3.89305562e-01 -1.53149441e-01 -3.12867880e-01
-7.43064433e-02 1.70493111e-01 -1.14782739e+00 2.07530022e+00
5.65545354e-03 7.01834679e-01 -2.97633737e-01 -9.63461041e-01
5.52618086e-01 1.78106114e-01 9.50434923e-01 -7.74297893e-01
2.20103487e-01 7.97666609e-02 -3.18177938e-01 -7.55284846e-01
1.80036247e-01 1.23566158e-01 2.11638212e-01 2.55759597e-01
2.24914644e-02 2.41202429e-01 1.81417957e-01 8.44323710e-02
1.16960263e+00 5.00533760e-01 3.87484282e-01 1.45275533e-01
5.99831581e-01 -2.52434075e-01 1.15517151e+00 4.62530106e-01
-7.79765069e-01 6.43243372e-01 1.09409221e-01 -6.17430449e-01
-7.52433062e-01 -1.02953672e+00 3.37518841e-01 9.97934103e-01
5.58910370e-01 -4.97637451e-01 -4.07063425e-01 -5.65262020e-01
-1.60242230e-01 1.49607241e-01 -5.91594160e-01 -4.17097420e-01
-7.13711321e-01 -2.54061073e-01 1.79317594e-01 7.72598267e-01
6.97526336e-01 -1.08477652e+00 -8.75277519e-01 1.60436243e-01
-4.93817985e-01 -1.18222296e+00 -1.07316136e+00 -4.29792851e-01
-8.48405898e-01 -1.01304650e+00 -8.73784542e-01 -6.14838123e-01
3.13850582e-01 8.22811425e-01 8.71248484e-01 6.65618060e-03
-4.17330593e-01 3.72380853e-01 -4.36869681e-01 2.29039177e-01
1.23950519e-01 -4.29428071e-01 4.34580110e-02 3.26110125e-01
4.67481226e-01 -7.90882945e-01 -1.05655146e+00 5.39393306e-01
-8.14916134e-01 1.39269456e-01 4.98882890e-01 9.68564987e-01
7.40823746e-01 -3.59746546e-01 2.69114316e-01 -1.99274704e-01
-9.43716615e-02 -3.65360290e-01 -3.50726157e-01 1.37180954e-01
-2.85409391e-01 -1.65551022e-01 3.20450217e-01 -5.81388593e-01
-8.77482831e-01 1.79959416e-01 1.22080214e-01 -1.00838470e+00
-9.21783000e-02 2.88100779e-01 -3.49016279e-01 5.74990455e-03
3.19582999e-01 4.58442897e-01 1.61074679e-02 -2.44060129e-01
3.14175546e-01 2.94927001e-01 7.04304218e-01 -2.84686685e-01
6.06845081e-01 8.00160110e-01 -9.90104675e-02 -6.62768722e-01
-1.00259495e+00 -7.09504128e-01 -6.86621606e-01 -6.82831168e-01
1.04641593e+00 -1.04539001e+00 -6.16104543e-01 8.28911006e-01
-8.79555106e-01 -6.48944139e-01 -3.82029772e-01 8.75709057e-01
-1.08332491e+00 6.40251756e-01 -7.40812719e-01 -4.93989557e-01
-1.64382488e-01 -9.82599080e-01 9.79238987e-01 2.47210532e-01
-5.20464480e-02 -8.18889260e-01 1.68376759e-01 4.55082923e-01
1.68248907e-01 2.97656715e-01 4.26219016e-01 2.76630744e-02
-8.58589828e-01 1.09435782e-01 -8.44547451e-02 3.07135314e-01
1.93590194e-01 -7.06368238e-02 -8.76702845e-01 -4.65293527e-01
4.15178761e-02 -5.03057003e-01 1.15116060e+00 6.14288568e-01
9.89782870e-01 -1.57471284e-01 -1.62097365e-01 9.45960939e-01
1.20424294e+00 1.20059356e-01 7.61369407e-01 2.47585654e-01
8.81941736e-01 3.72866333e-01 1.02874553e+00 6.91567838e-01
4.27783906e-01 9.02716339e-01 4.28459525e-01 1.65699452e-01
-4.47934777e-01 -2.78988242e-01 7.55200267e-01 8.52933884e-01
-2.94080704e-01 -2.19877549e-02 -5.67501664e-01 4.50102895e-01
-2.38167310e+00 -1.52215075e+00 1.62014082e-01 2.02982092e+00
5.41336358e-01 7.99435377e-02 3.36563081e-01 -9.51829273e-03
5.93337774e-01 8.60323310e-01 -7.88701296e-01 3.44602287e-01
-2.15570733e-01 -3.20735216e-01 9.43025947e-02 4.28674102e-01
-1.24161541e+00 9.02433395e-01 5.15345383e+00 8.80531132e-01
-1.04027045e+00 1.74534589e-01 5.58616579e-01 -4.47620839e-01
1.05938062e-01 6.15784898e-02 -6.48615956e-01 5.90916872e-01
4.29590970e-01 -1.25298813e-01 3.13662738e-01 6.85395300e-01
5.49879193e-01 -6.43869489e-02 -1.07309771e+00 1.29083121e+00
2.20188722e-01 -1.54994631e+00 -4.85401787e-02 5.22374734e-02
7.93600798e-01 6.28147647e-02 4.25397232e-02 1.51201814e-01
-1.28973827e-01 -7.20036030e-01 7.83764422e-01 8.82854044e-01
5.93533397e-01 -5.66080511e-01 3.69329661e-01 2.58628190e-01
-1.58198082e+00 -3.40187460e-01 -4.11127269e-01 -2.93271929e-01
4.19049621e-01 3.67491424e-01 1.78855941e-01 4.37863499e-01
7.15358078e-01 1.51258647e+00 -3.29777271e-01 1.05014050e+00
-1.01414330e-01 3.29325348e-01 1.49699807e-01 3.06874037e-01
4.57854748e-01 -2.08396018e-01 7.10460186e-01 8.45842898e-01
2.36708939e-01 2.89482713e-01 4.95495021e-01 6.73254192e-01
2.40043074e-01 -1.38064235e-01 -4.67087805e-01 7.26416409e-02
2.70469397e-01 9.88646865e-01 -3.02045852e-01 -3.66815776e-01
-7.84896910e-01 1.11946321e+00 2.94418603e-01 4.11526114e-01
-9.51876819e-01 -5.40128089e-02 1.04028332e+00 4.56976779e-02
4.51203942e-01 -3.62687290e-01 1.83921292e-01 -1.47740996e+00
1.97806120e-01 -8.64690900e-01 6.27587259e-01 -8.58983338e-01
-1.20229340e+00 2.90659994e-01 -1.18628688e-01 -1.72120869e+00
-2.55432636e-01 -3.90866667e-01 -7.36630917e-01 4.26686943e-01
-1.46637177e+00 -1.05082047e+00 -6.87372923e-01 8.79973352e-01
1.01227522e+00 -2.14730546e-01 4.03115779e-01 5.16651511e-01
-6.83008492e-01 4.45031106e-01 -5.90078756e-02 2.02503949e-01
8.49261105e-01 -7.33551621e-01 2.34538287e-01 9.47072327e-01
1.36447668e-01 1.42463118e-01 5.65882444e-01 -4.72587287e-01
-1.43837154e+00 -1.22315109e+00 6.35647893e-01 -2.91361630e-01
7.06872284e-01 -7.37003423e-03 -1.01199985e+00 5.62012970e-01
-1.15831748e-01 6.97917223e-01 3.71204257e-01 -5.57248294e-01
-2.92322576e-01 -1.65625617e-01 -5.64203680e-01 6.64736211e-01
1.42824900e+00 -7.25498557e-01 -5.42909622e-01 1.25800639e-01
5.69651544e-01 -2.65422642e-01 -7.85972357e-01 3.10649604e-01
7.58207977e-01 -1.15676379e+00 9.64987397e-01 -6.29533172e-01
6.55702472e-01 -4.20496643e-01 -4.48301435e-01 -8.40861142e-01
-4.72371668e-01 -7.47715354e-01 -7.19971299e-01 1.02497447e+00
-2.50228018e-01 -3.26411635e-01 7.88513303e-01 3.22064549e-01
-1.27792299e-01 -1.00483704e+00 -9.39278066e-01 -9.05874431e-01
-2.31658772e-01 -3.23119998e-01 2.76031226e-01 9.08769667e-01
-5.66124246e-02 2.30967611e-01 -8.11797500e-01 -1.17932089e-01
4.88245070e-01 2.34951720e-01 8.28475416e-01 -7.48615026e-01
-5.79206705e-01 -5.22048116e-01 -7.54495502e-01 -1.52936924e+00
2.91609377e-01 -4.99151081e-01 1.92304537e-01 -1.22456384e+00
4.34747338e-01 9.04671177e-02 -4.33059514e-01 3.02899182e-01
-4.18616831e-01 2.43874565e-01 3.64604503e-01 4.70135659e-01
-9.45203424e-01 1.03026378e+00 1.41806722e+00 -2.39205688e-01
-1.79243386e-01 -1.44326985e-01 -1.66259423e-01 9.29150462e-01
5.91957927e-01 -3.54153842e-01 -5.51727176e-01 -3.65880847e-01
-2.17053607e-01 3.39875698e-01 7.13409603e-01 -1.22188973e+00
3.08784813e-01 -5.95290482e-01 2.79343456e-01 -5.52132845e-01
6.28774345e-01 -4.89640713e-01 -3.41986641e-02 5.50067902e-01
-3.82713944e-01 6.04579784e-03 -3.50980669e-01 8.86500895e-01
-4.40364510e-01 1.44551933e-01 8.26534688e-01 -4.31071781e-02
-1.36394560e+00 8.71566594e-01 -2.67053366e-01 3.05931240e-01
1.07000780e+00 -3.58931214e-01 -2.98730701e-01 -5.16008794e-01
-5.10399818e-01 2.63041854e-01 5.61228991e-01 5.38202763e-01
8.31225574e-01 -1.70369375e+00 -5.87396264e-01 1.34496242e-01
2.98408091e-01 -2.38778770e-01 7.03980088e-01 1.23017299e+00
-2.41776198e-01 2.46414304e-01 -3.67996037e-01 -7.30174780e-01
-1.23079884e+00 6.34943724e-01 4.83842075e-01 -4.19186577e-02
-1.03855085e+00 7.36850262e-01 4.79728907e-01 2.04364881e-01
3.24658513e-01 -1.25225276e-01 -2.63846070e-02 -5.36730215e-02
7.96812475e-01 6.25186682e-01 -4.61499929e-01 -9.31256413e-01
-4.56891268e-01 8.00582469e-01 1.52414024e-01 2.93860864e-03
1.23083997e+00 -4.07911330e-01 2.75662512e-01 6.02604032e-01
1.37603688e+00 -5.63862681e-01 -2.01045084e+00 -4.87857074e-01
-2.68728256e-01 -1.05789268e+00 8.17405209e-02 -1.16021529e-01
-1.31578922e+00 1.00132918e+00 6.39022529e-01 -4.75767016e-01
1.28101408e+00 -1.33994609e-01 1.11151552e+00 3.67412984e-01
3.14143330e-01 -1.19594932e+00 7.48045802e-01 5.18542409e-01
7.25164890e-01 -1.44179010e+00 3.69442515e-02 -2.12455302e-01
-7.64980614e-01 1.00649571e+00 7.73235321e-01 -2.79801220e-01
6.58021033e-01 -1.29413471e-01 4.59128395e-02 -2.11478621e-01
-9.59381759e-01 -3.62261683e-01 3.89132977e-01 4.27824408e-01
1.92053482e-01 -3.40671122e-01 -2.13413730e-01 3.53805840e-01
4.83158708e-01 2.96174049e-01 1.39804110e-01 1.01141894e+00
-4.83757734e-01 -6.96700275e-01 4.28264029e-02 1.39798313e-01
-2.32598752e-01 1.73959836e-01 -6.96352720e-02 5.89069664e-01
-3.21715735e-02 8.69521260e-01 6.44696578e-02 -5.18265784e-01
1.29177853e-01 -2.26317316e-01 4.71313983e-01 -3.36267918e-01
-5.83653972e-02 2.17147440e-01 -1.11479081e-01 -1.35047901e+00
-7.32230484e-01 -7.84265399e-01 -1.18136287e+00 -3.73334140e-01
-4.31133062e-02 -3.20473135e-01 -8.97682533e-02 1.06355846e+00
3.72146785e-01 5.13163030e-01 7.77963459e-01 -1.09621918e+00
-6.16169274e-01 -6.34965122e-01 -5.92612684e-01 6.73022866e-01
4.73795146e-01 -9.11325336e-01 -3.43074083e-01 2.73183763e-01]
|
[8.635452270507812, 0.5151599645614624]
|
ab5fa383-157b-4ea8-a52d-0241670475c0
|
neural-document-summarization-by-jointly
|
1807.02305
| null |
http://arxiv.org/abs/1807.02305v1
|
http://arxiv.org/pdf/1807.02305v1.pdf
|
Neural Document Summarization by Jointly Learning to Score and Select Sentences
|
Sentence scoring and sentence selection are two main steps in extractive
document summarization systems. However, previous works treat them as two
separated subtasks. In this paper, we present a novel end-to-end neural network
framework for extractive document summarization by jointly learning to score
and select sentences. It first reads the document sentences with a hierarchical
encoder to obtain the representation of sentences. Then it builds the output
summary by extracting sentences one by one. Different from previous methods,
our approach integrates the selection strategy into the scoring model, which
directly predicts the relative importance given previously selected sentences.
Experiments on the CNN/Daily Mail dataset show that the proposed framework
significantly outperforms the state-of-the-art extractive summarization models.
|
['Nan Yang', 'Furu Wei', 'Shaohan Huang', 'Qingyu Zhou', 'Ming Zhou', 'Tiejun Zhao']
|
2018-07-06
|
neural-document-summarization-by-jointly-1
|
https://aclanthology.org/P18-1061
|
https://aclanthology.org/P18-1061.pdf
|
acl-2018-7
|
['extractive-document-summarization']
|
['natural-language-processing']
|
[ 5.71395397e-01 3.81273746e-01 -3.83155167e-01 -5.74653208e-01
-1.26468515e+00 -3.60710442e-01 3.78857017e-01 4.57491696e-01
-6.14382446e-01 9.35254574e-01 1.12522209e+00 1.47673100e-01
1.76185861e-01 -5.63588440e-01 -5.31538188e-01 -2.20059261e-01
3.96603405e-01 3.39253426e-01 1.52926311e-01 -1.83018297e-01
8.35724473e-01 1.25467246e-02 -1.24238551e+00 8.47405493e-01
1.16833913e+00 5.52320659e-01 3.80637139e-01 1.24363232e+00
-2.32632563e-01 1.27540255e+00 -1.21668577e+00 -4.12019998e-01
-2.57186234e-01 -8.03329468e-01 -1.09280062e+00 5.31121083e-02
8.11941743e-01 -6.63831532e-01 -3.41364682e-01 1.00438988e+00
8.55795860e-01 2.12171793e-01 6.16194725e-01 -4.47430134e-01
-8.38002086e-01 1.30133247e+00 -3.42093438e-01 4.55980241e-01
4.01290029e-01 -3.28111380e-01 1.49956357e+00 -9.08225000e-01
6.42387688e-01 1.01922774e+00 3.13285232e-01 7.55197287e-01
-8.85809004e-01 -2.31001258e-01 2.30749026e-01 4.34813127e-02
-6.62193000e-01 -7.33616710e-01 8.62547338e-01 6.66530356e-02
1.54785192e+00 5.06785512e-01 6.15720809e-01 8.81270766e-01
3.82840335e-01 1.51491201e+00 1.99773103e-01 -4.22184795e-01
1.60953522e-01 -3.17016691e-01 7.98826396e-01 5.45622349e-01
3.47084701e-01 -6.56397521e-01 -9.04487789e-01 -1.03500918e-01
1.22907743e-01 -1.04557581e-01 -1.78872064e-01 3.54846746e-01
-1.03977513e+00 8.21319401e-01 3.35707814e-01 2.13428736e-01
-7.76035905e-01 1.13408208e-01 8.45758498e-01 1.89665154e-01
9.59799707e-01 5.64183533e-01 -3.85062605e-01 4.06859517e-02
-1.71558321e+00 4.45295185e-01 9.74977016e-01 9.46133375e-01
3.18715900e-01 1.79354101e-01 -9.61028457e-01 9.49486375e-01
-2.53365450e-02 1.85809135e-01 6.60852671e-01 -7.94952393e-01
1.03772295e+00 5.76086938e-01 -1.03030019e-01 -6.70772672e-01
-3.36102217e-01 -6.18415594e-01 -1.05169833e+00 -4.51554984e-01
-4.52933013e-01 -4.73857969e-01 -8.25827062e-01 1.31656528e+00
-2.95390189e-01 -2.01154411e-01 2.89177984e-01 5.15472710e-01
1.59639311e+00 8.10458124e-01 -2.93587856e-02 -5.63774943e-01
1.12485087e+00 -1.51110673e+00 -1.01896918e+00 -3.52220416e-01
4.02344376e-01 -5.79269290e-01 6.81613564e-01 2.17355773e-01
-1.80839252e+00 -5.48346996e-01 -1.24689353e+00 -5.92861176e-01
1.82676733e-01 7.30269551e-01 2.12377831e-01 9.13437980e-04
-1.26005137e+00 7.73459792e-01 -8.30869973e-01 -3.12528700e-01
5.10466516e-01 3.35145801e-01 -8.38961378e-02 3.95270884e-01
-1.07360435e+00 8.78366649e-01 7.46532619e-01 1.62499696e-02
-5.26427984e-01 -3.59217167e-01 -8.66450667e-01 6.98341608e-01
2.16738805e-01 -1.33708358e+00 1.98294437e+00 -7.31270850e-01
-1.80706143e+00 5.11150539e-01 -7.28508353e-01 -7.92646468e-01
1.36516318e-01 -7.26624310e-01 1.97766453e-01 3.64342034e-01
2.85233885e-01 6.09983623e-01 6.55274570e-01 -9.39625084e-01
-7.46911108e-01 -3.84553611e-01 -8.98180380e-02 5.74343383e-01
-4.34604853e-01 4.14297879e-01 -3.58292162e-01 -6.11444294e-01
-2.11819142e-01 -3.07469964e-01 -3.07788014e-01 -1.01623619e+00
-1.03098595e+00 -4.78980094e-01 4.48816955e-01 -1.14573383e+00
1.75045133e+00 -1.46811593e+00 6.60306454e-01 -5.50277352e-01
2.53368646e-01 3.05419356e-01 -2.86280483e-01 7.74061859e-01
9.95702967e-02 5.31813502e-02 -3.77097666e-01 -1.03448784e+00
-4.17773835e-02 -3.48032981e-01 -4.95804608e-01 -2.18332648e-01
4.62179840e-01 1.08702433e+00 -9.52764690e-01 -8.21029246e-01
-1.77697897e-01 6.16858117e-02 -4.91857260e-01 3.47345442e-01
-2.10366532e-01 -8.05949494e-02 -5.70501804e-01 3.39977890e-01
3.92074585e-01 4.03822809e-02 2.93288101e-02 -5.93130384e-03
-1.76691756e-01 8.85801136e-01 -5.46028197e-01 1.81835282e+00
-2.06984743e-01 8.25562418e-01 -1.38066038e-01 -9.71594512e-01
8.25779974e-01 3.18736285e-01 1.71999354e-02 -4.13531840e-01
1.86993450e-01 1.24198988e-01 -3.33789557e-01 -7.12922454e-01
1.43420601e+00 3.35569680e-02 -3.51253688e-01 4.86589909e-01
5.24456203e-01 -3.46576631e-01 7.18576252e-01 7.40475237e-01
1.29974055e+00 -7.14405403e-02 7.07027555e-01 4.93284203e-02
5.93294740e-01 -7.54065737e-02 4.88540351e-01 9.10308659e-01
7.99880773e-02 1.06024206e+00 6.66287243e-01 -8.62108544e-02
-1.06823003e+00 -8.78476202e-01 4.61829513e-01 1.30533266e+00
-2.90868491e-01 -7.33944476e-01 -1.04376686e+00 -9.96090889e-01
-3.53731245e-01 1.18163240e+00 -4.32979524e-01 -2.01404184e-01
-9.71175313e-01 -5.00609636e-01 4.94055629e-01 8.13919067e-01
5.63414991e-01 -1.54237080e+00 -5.94790578e-01 4.17159200e-01
-5.85228443e-01 -7.65995145e-01 -7.36803174e-01 2.23124504e-01
-1.10993648e+00 -6.02480412e-01 -7.15958238e-01 -9.92357016e-01
5.70354283e-01 3.15229356e-01 1.35009110e+00 1.93426773e-01
1.97390899e-01 -5.71216084e-03 -4.46853071e-01 -7.63672471e-01
-5.89609087e-01 9.64605808e-01 -2.44110823e-01 -3.48882914e-01
3.18949431e-01 -4.49288338e-01 -6.17939472e-01 -7.63082743e-01
-9.79018152e-01 3.28581512e-01 1.02834940e+00 7.89225399e-01
3.87662590e-01 -3.59487444e-01 1.13805163e+00 -1.02183855e+00
1.49779534e+00 -1.69865146e-01 1.13624640e-01 3.80923510e-01
-1.90439776e-01 1.95710331e-01 9.02061045e-01 1.15882695e-01
-1.29262900e+00 7.24982983e-03 -2.67247915e-01 2.01539233e-01
-8.20783302e-02 9.14276004e-01 -1.89033151e-02 8.36417198e-01
4.40418184e-01 6.38220310e-01 -3.07069510e-01 -5.60990572e-01
2.91653395e-01 1.07805896e+00 6.39217079e-01 -2.37922408e-02
4.57285196e-01 1.02662809e-01 -5.42984724e-01 -9.54194129e-01
-1.55818236e+00 -5.47405064e-01 -9.71314728e-01 -7.67347589e-02
8.10473859e-01 -7.69189239e-01 -2.68120021e-01 2.94038087e-01
-1.81893492e+00 1.17048122e-01 -4.57725942e-01 1.61788583e-01
-5.44687629e-01 6.83892131e-01 -8.28084886e-01 -6.71867967e-01
-1.48046303e+00 -7.49032497e-01 1.47671175e+00 4.91469711e-01
-6.09932840e-01 -7.40416050e-01 2.58684158e-01 5.29009998e-01
3.08833033e-01 -1.29925177e-01 6.57645643e-01 -9.93309855e-01
-2.20309719e-01 -4.17969048e-01 -2.07501706e-02 7.23380029e-01
2.99263708e-02 1.19559728e-01 -6.32716537e-01 -1.39731228e-01
1.78917237e-02 -4.13577348e-01 1.85567439e+00 6.81753457e-01
1.09328854e+00 -6.77650571e-01 -6.60613030e-02 1.31486177e-01
9.72511113e-01 -1.24606237e-01 6.85565114e-01 2.07298517e-01
4.94475663e-01 5.70184171e-01 5.41058302e-01 4.45861399e-01
3.72715145e-01 1.48608074e-01 8.25675800e-02 1.11126781e-01
-1.26114488e-01 -2.06024066e-01 5.88142514e-01 1.51433134e+00
5.86354099e-02 -6.09411359e-01 -3.71582121e-01 6.29481316e-01
-2.19442487e+00 -1.42191815e+00 -3.13555375e-02 1.67374647e+00
1.16421962e+00 2.79005438e-01 8.75553582e-04 -4.16017510e-02
6.57192826e-01 7.31351256e-01 -5.54095924e-01 -7.76696086e-01
-1.47204727e-01 3.93102109e-01 1.93846032e-01 5.36594570e-01
-1.19766486e+00 1.02063334e+00 6.29290819e+00 6.39855564e-01
-7.61942267e-01 -1.57654867e-01 4.75652218e-01 -6.31421924e-01
-1.23564795e-01 -1.58311293e-01 -1.03479910e+00 3.64889860e-01
9.27559912e-01 -5.47043562e-01 -1.14037387e-01 6.49280012e-01
4.33814079e-01 -8.74247104e-02 -1.08403718e+00 3.24123472e-01
7.94366419e-01 -1.59593368e+00 5.74635625e-01 -4.82156396e-01
7.97011495e-01 9.77685302e-02 -3.22133303e-01 4.62905049e-01
2.35078290e-01 -7.55220652e-01 7.16224968e-01 7.62982607e-01
3.52760732e-01 -8.95707011e-01 1.05383670e+00 7.14945316e-01
-7.80185103e-01 2.05883794e-02 -5.69316328e-01 -2.17632547e-01
4.93730605e-01 6.01527333e-01 -8.56109083e-01 7.67925858e-01
2.63124526e-01 9.73402500e-01 -7.86164880e-01 1.04799938e+00
-5.99082351e-01 7.84058571e-01 2.52919585e-01 -4.62163568e-01
1.56667531e-01 -7.58404955e-02 8.00806105e-01 1.70322955e+00
2.49524042e-01 -2.68098526e-02 -6.43848069e-03 5.63442647e-01
-6.69158340e-01 1.63106412e-01 -3.27706814e-01 -2.85057444e-02
3.96506280e-01 1.25644708e+00 -4.22271878e-01 -8.56051803e-01
-1.43149510e-01 1.19230568e+00 7.40542948e-01 3.04122448e-01
-5.02672017e-01 -1.02949440e+00 -7.31201023e-02 -3.68176788e-01
4.51895326e-01 -8.61334801e-02 -5.35698891e-01 -1.54898119e+00
3.11350703e-01 -7.05651283e-01 2.44649187e-01 -7.66979814e-01
-9.38277185e-01 6.43061459e-01 -8.74635726e-02 -1.02530980e+00
-4.64923233e-01 -3.07230698e-03 -1.12720823e+00 6.21030629e-01
-1.56685269e+00 -1.05115390e+00 -1.63039923e-01 -2.41367027e-01
1.57742274e+00 -2.08368331e-01 7.46817291e-01 -2.36412257e-01
-8.37779939e-01 2.60293871e-01 2.51129240e-01 2.36428186e-01
5.92178643e-01 -1.52841616e+00 7.45388627e-01 1.03287160e+00
-4.38806117e-02 7.04929292e-01 7.36752152e-01 -8.29452693e-01
-8.06517124e-01 -1.17138135e+00 1.55459428e+00 -2.50655085e-01
2.32924983e-01 -2.08230898e-01 -7.52277911e-01 7.08921671e-01
1.21631432e+00 -9.64250743e-01 7.39209771e-01 1.68744717e-02
1.46491736e-01 1.44727781e-01 -6.52336121e-01 7.60750353e-01
9.27306712e-01 -1.78891256e-01 -1.30880368e+00 3.79739106e-01
1.14974678e+00 -3.91421437e-01 -3.80765557e-01 2.71845222e-01
4.07148063e-01 -8.47022712e-01 7.04936862e-01 -9.08953667e-01
1.24011409e+00 1.51861712e-01 2.75333166e-01 -1.80149412e+00
-4.72518265e-01 -5.22044539e-01 -5.37021756e-01 1.36712646e+00
5.98768413e-01 -1.45085722e-01 7.59486973e-01 1.37165681e-01
-6.47781789e-01 -8.19599271e-01 -6.89970255e-01 -3.87200266e-01
1.12010933e-01 1.36886448e-01 4.39856976e-01 2.99471408e-01
2.79352695e-01 1.24653423e+00 -3.26923996e-01 -3.43449891e-01
5.48625529e-01 2.10566282e-01 6.60089254e-01 -1.19509411e+00
-1.19813301e-01 -8.23185623e-01 1.90433368e-01 -1.31704366e+00
5.44936836e-01 -1.04789901e+00 3.43422443e-01 -2.44117379e+00
8.68143678e-01 9.01335418e-01 -1.39892220e-01 2.62403697e-01
-7.59376466e-01 -2.35823348e-01 3.31424356e-01 8.46204907e-02
-1.14029849e+00 9.51951504e-01 9.94649291e-01 -5.20149350e-01
-3.83829117e-01 2.58843005e-01 -1.17190325e+00 6.44712985e-01
9.84621525e-01 -4.77267355e-01 -1.67791963e-01 -7.40066946e-01
2.72489488e-01 2.58105099e-01 -7.18925446e-02 -9.50986087e-01
5.61870694e-01 4.39617261e-02 3.86226267e-01 -1.49187660e+00
1.12907566e-01 1.50932875e-02 -8.47813249e-01 3.14353794e-01
-1.21859276e+00 3.97089273e-02 -1.84579846e-02 4.63733613e-01
-4.97293949e-01 -7.76804864e-01 4.13907826e-01 -2.49254853e-01
-1.09160356e-01 -7.83619508e-02 -6.16047084e-01 8.48593414e-02
3.91928941e-01 -2.39652190e-02 -4.84322667e-01 -5.82476735e-01
-5.03117144e-01 4.95320022e-01 -7.43330345e-02 3.26305002e-01
1.03730774e+00 -9.51219916e-01 -1.44485462e+00 -2.51373678e-01
-2.21924216e-01 4.01798457e-01 2.01763526e-01 5.56897581e-01
-4.39200222e-01 7.56950617e-01 8.85047019e-02 -1.05070427e-01
-1.75655425e+00 7.47042969e-02 -6.88462034e-02 -7.75970757e-01
-4.72640008e-01 7.72483289e-01 -7.37383291e-02 -2.71242380e-01
2.78346241e-01 -3.78197461e-01 -7.49085307e-01 3.83148253e-01
8.21858704e-01 5.41787505e-01 2.06929237e-01 -3.36643785e-01
7.36257955e-02 1.33826777e-01 -7.33642161e-01 -1.50141954e-01
1.80027604e+00 -1.32375821e-01 -4.67838287e-01 2.81336546e-01
1.24736929e+00 1.84937529e-02 -7.35674500e-01 -2.71584541e-01
2.79880226e-01 9.01744142e-02 -2.83798072e-02 -7.73673713e-01
-6.60570920e-01 8.55180562e-01 -4.65215951e-01 3.76568705e-01
1.22683549e+00 2.06797700e-02 1.26687324e+00 8.89715374e-01
-3.51999581e-01 -1.49090743e+00 3.15742970e-01 1.05340469e+00
1.32031000e+00 -1.00971615e+00 3.63709629e-01 -9.78293195e-02
-8.25104892e-01 1.36454248e+00 6.32816255e-01 -4.17522341e-01
-1.15939669e-01 1.80925056e-02 -3.87971640e-01 -2.06246600e-01
-1.17121601e+00 -1.01131000e-01 4.04166579e-01 1.22878760e-01
8.66759300e-01 -6.47288412e-02 -7.66283214e-01 9.82072473e-01
-5.68221509e-01 8.89813378e-02 9.05629456e-01 1.04434121e+00
-1.00561941e+00 -8.24472368e-01 5.07479534e-02 8.88752103e-01
-6.68040931e-01 -4.35414076e-01 -1.12494123e+00 1.92003131e-01
-6.81885898e-01 1.06664872e+00 5.53432405e-02 -2.57678896e-01
6.51270926e-01 2.76534945e-01 2.49971703e-01 -1.22498834e+00
-1.16965222e+00 -8.99536386e-02 4.60128069e-01 -6.90942928e-02
-3.52277607e-01 -7.55014598e-01 -1.35704768e+00 6.23833314e-02
-4.52699184e-01 2.91044176e-01 4.93549824e-01 7.91084290e-01
7.12107360e-01 1.07251596e+00 7.32715428e-01 -1.11713815e+00
-1.03398502e+00 -1.56954217e+00 -2.55513728e-01 1.80743203e-01
5.88711381e-01 2.70389229e-01 -2.14522123e-01 1.33402869e-01]
|
[12.603411674499512, 9.394997596740723]
|
975a02be-ef56-4cda-b151-a56fe538b24f
|
sleep-stage-classification-from-heart-rate
| null | null |
https://doi.org/10.1038/s41598-019-49703-y
|
https://www.nature.com/articles/s41598-019-49703-y.pdf
|
Sleep stage classification from heart-rate variability using long short-term memory neural networks
|
Automated sleep stage classification using heart rate variability (HRV) may provide an ergonomic and low-cost alternative to gold standard polysomnography, creating possibilities for unobtrusive home-based sleep monitoring. Current methods however are limited in their ability to take into account long-term sleep architectural patterns. A long short-term memory (LSTM) network is proposed as a solution to model long-term cardiac sleep architecture information and validated on a comprehensive data set (292 participants, 584 nights, 541.214 annotated 30 s sleep segments) comprising a wide range of ages and pathological profiles, annotated according to the Rechtschaffen and Kales (R&K) annotation standard. It is shown that the model outperforms state-of-the-art approaches which were often limited to non-temporal or short-term recurrent classifiers. The model achieves a Cohen’s k of 0.61 ± 0.15 and accuracy of 77.00 ± 8.90% across the entire database. Further analysis revealed that the performance for individuals aged 50 years and older may decline. These results demonstrate the merit of deep temporal modelling using a diverse data set and advance the state-of-the-art for HRV-based sleep stage classification. Further research is warranted into individuals over the age of 50 as performance tends to worsen in this sub-population.
|
['Arnaud Moreau', 'Mustafa Radha', 'Ronald M. Aarts', 'Peter Anderer', 'Pedro Fonseca', 'Xi Long', 'Marco Ross', 'Andreas Cerny']
|
2019-10-02
| null | null | null |
scientific-reports-2019-10
|
['sleep-stage-detection', 'heart-rate-variability', 'electrocardiography-ecg']
|
['medical', 'medical', 'methodology']
|
[-1.23061694e-01 -1.94103226e-01 -2.77818471e-01 -3.77043784e-01
-2.81225801e-01 -9.68575701e-02 5.65722696e-02 -1.76557805e-02
-6.86681509e-01 7.89447784e-01 2.13043302e-01 -1.78866580e-01
-1.20959453e-01 -4.46771294e-01 2.94737160e-01 -5.78862309e-01
-4.19467390e-01 1.56457007e-01 -2.55544960e-01 -2.24281728e-01
-1.80137902e-01 -5.52030392e-02 -1.63719058e+00 -1.34382561e-01
6.33457363e-01 1.28483856e+00 -4.77537215e-02 6.71546459e-01
4.63018328e-01 1.20168664e-01 -8.85226548e-01 -8.22757334e-02
-7.96363801e-02 -5.78730524e-01 -4.97640461e-01 -2.97229618e-01
2.76984602e-01 1.64173558e-01 -9.34104398e-02 4.96787846e-01
1.06060672e+00 3.95898849e-01 -1.35819959e-02 -7.39720464e-01
-6.08062387e-01 -3.24114375e-02 2.75602136e-02 1.12920833e+00
4.51746315e-01 2.31777921e-01 8.00942779e-01 -3.67775470e-01
2.13221282e-01 3.96914661e-01 1.29933870e+00 8.31496239e-01
-1.27889800e+00 -8.10688972e-01 -3.28299999e-01 4.53585654e-01
-1.52245438e+00 -7.53996730e-01 5.56032896e-01 -2.47179881e-01
1.54385304e+00 4.21984613e-01 1.54860818e+00 1.33557546e+00
8.56631815e-01 -2.03270152e-01 1.27160680e+00 -8.29916224e-02
2.13425145e-01 -1.33964002e-01 5.87177932e-01 5.41276097e-01
3.97537023e-01 3.36121380e-01 -8.03790629e-01 1.72631275e-02
1.05569556e-01 2.54922479e-01 1.40063509e-01 4.44916129e-01
-8.88207853e-01 4.81032401e-01 3.09621722e-01 7.61255026e-01
-4.07985270e-01 -1.04585588e-01 6.76631510e-01 4.76427495e-01
8.60069692e-01 3.30624700e-01 -6.17574871e-01 -7.00642645e-01
-1.67685997e+00 3.77033763e-02 7.06933618e-01 2.26766348e-01
2.62356520e-01 1.98629200e-01 -1.14274472e-01 8.90369713e-01
2.72078514e-01 4.01080191e-01 1.03943062e+00 -1.05099607e+00
-1.15316331e-01 6.85496986e-01 -1.82333872e-01 -6.78812206e-01
-1.38840520e+00 -9.67829406e-01 -1.17929435e+00 -1.83418304e-01
1.26604319e-01 1.38430685e-01 -6.46789670e-01 1.56930792e+00
-3.01407695e-01 1.54949233e-01 -1.85597256e-01 6.71280265e-01
1.02418327e+00 1.31915838e-01 9.08179581e-02 -7.59391308e-01
1.76996791e+00 -5.50687730e-01 -8.26196015e-01 -6.44719183e-01
1.29117802e-01 -1.60158828e-01 9.79112744e-01 5.30445158e-01
-1.10904729e+00 -8.84679258e-01 -1.22347450e+00 -1.45903036e-01
-3.31422448e-01 5.92733994e-02 3.47091168e-01 1.30271649e+00
-1.34574544e+00 9.09286857e-01 -1.29662204e+00 -7.89392829e-01
3.96977127e-01 7.54742026e-01 3.00084576e-02 6.18850410e-01
-1.21632826e+00 9.25609589e-01 -1.24320515e-01 5.98944068e-01
-6.09032154e-01 -7.05706894e-01 -7.33793616e-01 -4.74499986e-02
-2.20272988e-01 -1.19936645e+00 9.32530820e-01 -5.31981230e-01
-1.36805856e+00 1.18572211e+00 -5.71464598e-01 -7.77610242e-01
9.68011767e-02 -1.38348922e-01 -9.31106925e-01 6.65808916e-02
1.35366842e-01 2.27432728e-01 7.31933057e-01 -2.33463556e-01
-5.99992052e-02 -7.27374732e-01 -5.05256474e-01 -1.91226900e-01
-3.08525234e-01 1.15541436e-01 2.42648572e-01 -4.80099112e-01
2.24155784e-02 -1.17671442e+00 -1.31515250e-01 -3.23167324e-01
-5.07155992e-02 -1.42821401e-01 3.45132440e-01 -8.44980061e-01
1.77100480e+00 -1.82330120e+00 -1.96920782e-01 -1.01956710e-01
5.77704489e-01 2.86768585e-01 5.61046600e-01 1.71052590e-01
-1.14240602e-01 2.48438314e-01 -3.14545818e-02 -9.99891579e-01
-1.14536822e-01 4.93699610e-01 3.25151086e-01 8.03947031e-01
-4.95639563e-01 9.97334719e-01 -6.29372180e-01 -3.11474621e-01
4.03443724e-01 5.33882201e-01 -1.10934466e-01 -4.69146483e-02
5.25023043e-01 4.49448526e-01 2.11068109e-01 6.94931030e-01
9.85543877e-02 -3.58398497e-01 -4.63920385e-02 1.04853369e-01
-3.18983167e-01 5.29107153e-01 -3.60542923e-01 1.96599221e+00
-6.25970244e-01 9.38953578e-01 -3.97220910e-01 -6.33774817e-01
9.56188858e-01 4.56222773e-01 4.96301204e-01 -9.14470553e-01
4.41006064e-01 1.48049742e-01 3.91127169e-01 -5.34068882e-01
2.97925264e-01 -7.75659144e-01 -1.71581715e-01 3.57778817e-01
8.75575021e-02 5.95832109e-01 1.20012745e-01 -5.08044422e-01
1.29819310e+00 -3.15383404e-01 5.86796463e-01 -4.60048258e-01
5.24216771e-01 -6.06774569e-01 8.16622674e-01 6.36295259e-01
-7.98201799e-01 5.07898867e-01 1.13091990e-01 -7.47648478e-01
-5.98256111e-01 -9.10513759e-01 -3.22555214e-01 8.95069599e-01
-3.80102277e-01 -9.09964085e-01 -5.03387034e-01 -2.27133587e-01
-3.54320794e-01 4.73963261e-01 -9.90115404e-01 -6.12498045e-01
-2.68756121e-01 -1.11740828e+00 8.93190682e-01 5.20316422e-01
5.79236448e-01 -1.30216384e+00 -1.10237992e+00 2.55629659e-01
-3.56986284e-01 -9.21447635e-01 -2.39441216e-01 2.32243672e-01
-1.12799084e+00 -9.30139542e-01 -6.02942944e-01 -4.13942039e-01
1.20268866e-01 -1.01395138e-01 1.37115896e+00 5.64955212e-02
-4.96553510e-01 3.13543081e-01 -1.08545162e-01 -2.67006129e-01
-2.20494289e-02 4.26299989e-01 5.92734933e-01 -2.76366204e-01
7.97822952e-01 -1.18709350e+00 -1.19102025e+00 3.46100539e-01
6.46281913e-02 -4.51138377e-01 4.22392309e-01 6.53904855e-01
4.78055209e-01 -7.35298619e-02 7.59529948e-01 -2.30551854e-01
5.55664420e-01 -3.70739341e-01 1.00676790e-02 -2.20860720e-01
-1.45683789e+00 -2.62847334e-01 5.26858747e-01 -3.31417739e-01
-6.07639015e-01 -5.31673133e-01 -3.31404030e-01 -2.57407218e-01
-1.23805679e-01 2.47720495e-01 3.41831803e-01 2.05306575e-01
6.80861950e-01 4.02680457e-01 2.75399655e-01 -5.13174474e-01
-5.00035167e-01 6.18322611e-01 5.96740186e-01 7.50533268e-02
2.82028705e-01 4.00205612e-01 1.83383316e-01 -1.36477494e+00
-9.22093928e-01 -6.64487720e-01 -7.25439191e-01 -2.53366977e-01
1.19472826e+00 -9.32232022e-01 -9.88422930e-01 3.83455276e-01
-3.33001077e-01 -5.30596137e-01 -3.50568235e-01 3.44622672e-01
-4.77797419e-01 1.32494137e-01 -4.78666335e-01 -1.05913138e+00
-1.19787526e+00 -5.46950638e-01 8.19900155e-01 2.98052728e-01
-9.89222109e-01 -1.06427658e+00 4.58993882e-01 5.85689306e-01
5.85442364e-01 3.52204859e-01 3.31030399e-01 -5.40533543e-01
3.10322553e-01 -1.08869173e-01 2.97976375e-01 4.22073007e-01
2.56732285e-01 -4.86596406e-01 -1.02334249e+00 -2.57189840e-01
5.13151705e-01 6.57023713e-02 7.61495411e-01 7.26113856e-01
8.21877122e-01 -8.71981829e-02 -1.18936762e-01 6.16823852e-01
1.13147080e+00 1.34524912e-01 8.19529593e-01 5.93746006e-01
4.23812330e-01 1.80614233e-01 4.58357558e-02 3.52053881e-01
6.69263184e-01 4.67827529e-01 1.49249211e-01 1.62649572e-01
-1.26751617e-01 2.78759122e-01 5.38489521e-01 6.49074972e-01
-4.52784240e-01 3.57053190e-01 -7.86257625e-01 4.18787450e-01
-1.41904104e+00 -1.28398728e+00 -1.82815030e-01 2.33025599e+00
4.48972672e-01 5.19000113e-01 7.65724182e-01 4.54210639e-01
3.49711388e-01 4.14599568e-01 -5.83460987e-01 -6.52599871e-01
2.00486198e-01 7.07339764e-01 2.57570952e-01 -1.19155377e-01
-7.86685228e-01 2.77379513e-01 6.62941027e+00 1.64513484e-01
-1.08213973e+00 5.16129434e-01 4.70297039e-01 -7.30978072e-01
5.35996735e-01 -3.25795650e-01 -7.02828169e-01 9.05416012e-01
2.24339414e+00 -7.09325448e-02 6.64779961e-01 5.19434988e-01
9.59226310e-01 -2.62420237e-01 -7.47101307e-01 1.31705987e+00
3.40881467e-01 -9.97290671e-01 -1.13646030e+00 3.61440629e-01
2.76534498e-01 3.48791063e-01 -6.47330657e-02 5.59103072e-01
-8.97503436e-01 -1.03864479e+00 3.69920224e-01 9.15285885e-01
1.03524506e+00 -4.73095179e-01 8.23498428e-01 1.25595510e-01
-1.38710558e+00 -3.40419501e-01 -3.31129432e-02 -6.93411052e-01
2.21747190e-01 4.00405943e-01 -4.37990636e-01 2.76403368e-01
1.16525340e+00 1.16222203e+00 -1.12773454e+00 8.56886744e-01
1.73559472e-01 7.87693560e-01 -3.68196249e-01 -1.07292533e-01
-2.35287309e-01 -5.43174669e-02 2.80439258e-01 9.76348639e-01
4.15837169e-01 -2.01733410e-02 -3.20255429e-01 7.54229963e-01
2.03842357e-01 -4.02427197e-01 -4.92784530e-01 -2.72082929e-02
6.19741976e-02 1.32825100e+00 -8.81244421e-01 -1.66295841e-01
-5.69345295e-01 7.97094882e-01 -2.20372632e-01 -7.64593249e-03
-6.24282479e-01 -1.63520455e-01 6.08914673e-01 3.37981075e-01
-7.45529458e-02 -2.54551053e-01 -7.96640992e-01 -9.72968578e-01
3.85932699e-02 -5.68730652e-01 6.49174631e-01 -6.27638638e-01
-1.19495535e+00 6.41185224e-01 -1.09987170e-01 -1.26915908e+00
-1.39635831e-01 -1.51565000e-01 -8.40116322e-01 7.33921289e-01
-1.07546771e+00 -1.09463191e+00 -5.42259991e-01 3.05175602e-01
6.66117251e-01 -3.52156311e-02 9.80750322e-01 4.99259263e-01
-1.02785027e+00 5.48858762e-01 -3.35815609e-01 -3.38621736e-01
5.05530059e-01 -1.30517030e+00 3.26175332e-01 7.24294066e-01
-2.73044139e-01 1.11500084e+00 8.18425894e-01 -4.52983707e-01
-1.00600040e+00 -1.03503799e+00 1.20595145e+00 -6.49299383e-01
4.74325895e-01 -2.82571673e-01 -5.24399877e-01 5.47926903e-01
1.66196868e-01 -9.95340347e-02 1.41900480e+00 6.31088674e-01
3.86418134e-01 -5.93541563e-01 -1.06614077e+00 1.47842497e-01
1.12201107e+00 -6.28235221e-01 -8.59170854e-01 -1.72305167e-01
1.64370954e-01 -9.56003368e-02 -1.19329083e+00 2.95741946e-01
1.21237409e+00 -1.39306533e+00 9.41065371e-01 -1.19081112e-02
-1.48076996e-01 1.32076964e-01 2.06257686e-01 -7.99575865e-01
-2.41633371e-01 -9.50697124e-01 -5.08972824e-01 9.57419276e-01
7.41239116e-02 -6.58085048e-01 8.53249252e-01 7.60094285e-01
-5.46196401e-01 -8.24498773e-01 -1.37013853e+00 -8.70949209e-01
-2.99575359e-01 -6.82146311e-01 1.72142833e-01 3.46794099e-01
3.69372107e-02 6.28247857e-01 -5.00425279e-01 -1.76540494e-01
3.97122443e-01 -9.11583230e-02 2.15375409e-01 -1.70709121e+00
8.36226270e-02 -3.83161575e-01 -6.01713002e-01 -1.22594245e-01
2.33646154e-01 -8.00215960e-01 -2.97012627e-01 -1.62212694e+00
1.04122221e-01 -3.00865285e-02 -7.41415083e-01 4.71816689e-01
1.10730350e-01 7.13549078e-01 -2.75584042e-01 1.50992110e-01
-7.14803398e-01 5.10223746e-01 6.66262507e-01 1.16789818e-01
-7.97807157e-01 5.40219545e-01 -8.30939472e-01 5.76848328e-01
1.17582238e+00 -4.47337806e-01 -5.31088054e-01 4.03285325e-01
3.92232895e-01 1.06536962e-01 4.97215629e-01 -1.54614174e+00
3.69142257e-02 6.11162484e-01 7.68095255e-01 -4.09779310e-01
7.71788657e-01 -5.10228693e-01 4.41424042e-01 8.65443349e-01
-1.52297709e-02 4.98477548e-01 1.09885678e-01 5.09286582e-01
3.25010449e-01 1.79669157e-01 6.65808618e-01 -3.21423203e-01
-3.60180080e-01 1.78776249e-01 -7.39551067e-01 4.23597880e-02
5.00110745e-01 -6.36148751e-01 -3.99110377e-01 -2.06090271e-01
-1.40880644e+00 -6.56555733e-03 3.38603437e-01 4.69472677e-01
4.10016507e-01 -1.03751540e+00 -5.84338233e-02 4.83734250e-01
1.38963178e-01 -5.83822668e-01 5.99978626e-01 1.48155069e+00
-1.84201390e-01 6.42317593e-01 -4.14717078e-01 -5.40458918e-01
-1.48185825e+00 4.03916568e-01 6.61122143e-01 -3.05708200e-01
-9.80578482e-01 5.65999031e-01 -7.55773902e-01 1.40868038e-01
3.66940089e-02 -6.89818859e-01 -4.19794053e-01 5.23404241e-01
5.78802109e-01 1.03534198e+00 4.44731027e-01 -7.92940915e-01
-6.99668884e-01 4.18584704e-01 3.88457030e-01 1.27911970e-01
1.35903788e+00 -6.15682304e-01 -3.03783774e-01 1.14493310e+00
1.05734909e+00 -3.32653403e-01 -6.23797834e-01 4.68235821e-01
-3.17069003e-03 8.47878829e-02 2.19581798e-02 -7.08921909e-01
-7.96619952e-01 7.48295605e-01 1.53860998e+00 4.56505775e-01
1.30459762e+00 -1.60151452e-01 1.14586997e+00 3.36389422e-01
1.84982166e-01 -1.02394712e+00 -1.65142432e-01 8.77555236e-02
4.08123255e-01 -9.26683128e-01 2.78251469e-01 5.51478028e-01
-3.69366258e-01 1.08703220e+00 3.82780969e-01 -1.49230629e-01
8.13465297e-01 -4.17200536e-01 2.00954149e-03 -5.16687691e-01
-7.76560307e-01 -4.00792152e-01 4.13231134e-01 5.84410667e-01
4.19426948e-01 6.04654998e-02 -6.90819561e-01 1.02568197e+00
-9.09271896e-01 2.29671001e-01 4.98451769e-01 6.57296717e-01
-2.72144794e-01 -1.07278514e+00 -4.38756496e-02 1.08290768e+00
-1.00665057e+00 1.05441827e-02 -1.16701694e-02 6.45119131e-01
4.60850805e-01 1.39488637e+00 1.09013952e-01 -5.38924277e-01
2.62412250e-01 6.03799641e-01 3.11951458e-01 -9.60141957e-01
-9.38801229e-01 -1.38981268e-02 3.19021851e-01 -5.88424623e-01
-7.85018325e-01 -9.99593437e-01 -8.67698491e-01 -1.62116840e-01
1.24085464e-01 5.63964583e-02 3.95654172e-01 1.10541427e+00
5.73866069e-01 7.87673652e-01 1.80086270e-01 -9.93303657e-01
7.84721971e-02 -1.22428107e+00 -6.54463112e-01 -1.13559447e-01
8.48704100e-01 -7.16096044e-01 -4.33294237e-01 -1.43103581e-02]
|
[13.554085731506348, 3.462825059890747]
|
3b059bc6-54d7-4a04-8119-064f92f00a57
|
a-pilot-study-on-dialogue-level-dependency
|
2305.12441
| null |
https://arxiv.org/abs/2305.12441v2
|
https://arxiv.org/pdf/2305.12441v2.pdf
|
A Pilot Study on Dialogue-Level Dependency Parsing for Chinese
|
Dialogue-level dependency parsing has received insufficient attention, especially for Chinese. To this end, we draw on ideas from syntactic dependency and rhetorical structure theory (RST), developing a high-quality human-annotated corpus, which contains 850 dialogues and 199,803 dependencies. Considering that such tasks suffer from high annotation costs, we investigate zero-shot and few-shot scenarios. Based on an existing syntactic treebank, we adopt a signal-based method to transform seen syntactic dependencies into unseen ones between elementary discourse units (EDUs), where the signals are detected by masked language modeling. Besides, we apply single-view and multi-view data selection to access reliable pseudo-labeled instances. Experimental results show the effectiveness of these baselines. Moreover, we discuss several crucial points about our dataset and approach.
|
['Min Zhang', 'Meishan Zhang', 'Shuang Liu', 'Gongyao Jiang']
|
2023-05-21
| null | null | null | null |
['dependency-parsing']
|
['natural-language-processing']
|
[ 1.22368664e-01 5.18666804e-01 -2.80921757e-01 -6.03583395e-01
-1.22473896e+00 -8.12882304e-01 6.96047246e-01 -1.19675025e-01
-3.72812629e-01 1.02786720e+00 7.59633839e-01 -1.99894175e-01
5.47957242e-01 -4.54825073e-01 -4.51114655e-01 -4.75436926e-01
-2.64807176e-02 2.36896396e-01 3.63511294e-01 -3.50431383e-01
1.67174578e-01 -1.72705129e-01 -1.02325630e+00 7.00987756e-01
8.87628376e-01 2.43632242e-01 2.23540738e-01 7.99914420e-01
-4.50378686e-01 1.17553294e+00 -1.04113424e+00 -8.62502038e-01
-2.57316023e-01 -8.93277347e-01 -1.06877029e+00 6.39750510e-02
-2.02771038e-01 -3.19026887e-01 -1.93278328e-01 1.04720569e+00
6.32915735e-01 -3.60254161e-02 3.96338850e-01 -6.69031322e-01
-6.86945260e-01 9.92700219e-01 -5.59494197e-01 4.41287249e-01
5.40138364e-01 7.17161819e-02 1.49868751e+00 -1.16756821e+00
9.58215892e-01 1.45697987e+00 2.08413646e-01 7.35090673e-01
-9.48062003e-01 -2.70876557e-01 4.62071419e-01 1.00307047e-01
-6.97551548e-01 -7.89566398e-01 8.35167468e-01 -3.74969155e-01
1.04959452e+00 9.85497385e-02 1.71993673e-01 1.69755757e+00
-4.21063267e-02 1.11056268e+00 1.17038202e+00 -9.05988216e-01
1.42597079e-01 4.32276353e-02 5.74024856e-01 5.54542422e-01
-9.43077579e-02 -3.09140593e-01 -7.72350907e-01 9.00938734e-03
2.96784312e-01 -6.72526062e-01 -3.70780200e-01 3.71576339e-01
-1.16136336e+00 9.28728759e-01 -7.11158738e-02 7.09136069e-01
-5.74827194e-02 -4.02203560e-01 7.32799411e-01 3.65954429e-01
7.58921146e-01 3.25060934e-01 -4.62829024e-01 -2.34995872e-01
-4.50840235e-01 -2.64192641e-01 9.94175017e-01 1.33709776e+00
3.89776945e-01 -1.42910495e-01 -6.73479438e-01 8.58060539e-01
1.74819320e-01 2.01511860e-01 3.54688555e-01 -8.46641839e-01
1.06941593e+00 4.82419670e-01 9.47369933e-02 -5.87627292e-01
-5.48096299e-01 2.06197262e-01 -5.68572819e-01 -5.37287891e-01
5.52489936e-01 -6.79433405e-01 -4.21085596e-01 1.61487806e+00
4.61408973e-01 -6.54243380e-02 4.10786331e-01 9.05278981e-01
1.08472979e+00 6.09681547e-01 1.57127082e-01 -6.54398143e-01
1.71085203e+00 -1.17693388e+00 -1.26908731e+00 -3.82647812e-01
1.03559625e+00 -7.28366494e-01 1.22821236e+00 8.33511576e-02
-1.00241542e+00 -3.95913333e-01 -9.36513245e-01 -3.20813656e-01
1.82851534e-02 1.94337085e-01 4.92828459e-01 5.97016931e-01
-4.68833357e-01 2.51129836e-01 -7.65023172e-01 -7.65208304e-02
4.39609103e-02 -2.44578347e-01 -1.31734967e-01 -6.93067722e-03
-1.63832700e+00 8.99268091e-01 4.10930991e-01 2.37896100e-01
-7.07934797e-01 -2.47291386e-01 -1.15316033e+00 -1.34642348e-01
8.22918415e-01 -1.81570336e-01 1.56219816e+00 -8.13670814e-01
-1.88121295e+00 1.04166543e+00 -4.58022743e-01 -3.34141016e-01
2.00864583e-01 -4.61742908e-01 -3.86927903e-01 2.01655746e-01
2.86890179e-01 7.74916336e-02 4.38678324e-01 -1.04694545e+00
-6.48928165e-01 -2.61166871e-01 3.21749926e-01 3.88141572e-01
-2.21487716e-01 6.53972566e-01 -6.95738018e-01 -5.61859667e-01
-1.32686734e-01 -7.14721918e-01 -1.96690664e-01 -8.42619181e-01
-9.62474525e-01 -4.52519804e-01 4.34821039e-01 -8.08667719e-01
1.40580773e+00 -2.16038871e+00 3.13143522e-01 -4.31227535e-01
9.56480280e-02 1.30622908e-01 -1.19218662e-01 6.25268817e-01
1.79684654e-01 1.98292002e-01 -2.89801508e-01 -5.58787107e-01
-8.52244049e-02 4.86238539e-01 -2.93621123e-01 2.67293870e-01
4.83463347e-01 1.00161242e+00 -1.09655690e+00 -7.20176280e-01
5.71407750e-02 5.68927638e-02 -2.73618668e-01 6.26141846e-01
-4.02779818e-01 8.86326134e-01 -7.71009624e-01 7.49230742e-01
2.06586123e-01 -1.34979144e-01 7.09531963e-01 -1.39354855e-01
-3.35071564e-01 9.23063993e-01 -6.13183081e-01 1.91568196e+00
-4.00229394e-01 5.46323776e-01 2.37419084e-01 -9.39073503e-01
8.95785987e-01 4.70289588e-01 -1.10650480e-01 -6.11310005e-01
2.11277470e-01 4.82011726e-03 9.71009210e-02 -8.96146774e-01
4.49347079e-01 -1.79349184e-01 -5.70319235e-01 3.70512664e-01
3.31569195e-01 1.85313493e-01 5.60935736e-01 4.67787236e-01
1.21515298e+00 2.56774426e-01 4.99676704e-01 -7.37216696e-02
6.03818953e-01 1.52218789e-01 7.96424747e-01 6.46148026e-01
-2.59708583e-01 5.90063691e-01 9.31172371e-01 -2.39757020e-02
-6.17718518e-01 -6.43038571e-01 1.03849620e-01 1.39296639e+00
3.47857289e-02 -5.38384318e-01 -9.20972705e-01 -1.05908144e+00
-6.76817417e-01 9.71987247e-01 -4.38586533e-01 2.88758993e-01
-1.08151138e+00 -1.05879486e+00 6.51798010e-01 4.25875932e-01
4.67661858e-01 -1.28962040e+00 -5.62962115e-01 5.03411591e-01
-7.38404095e-01 -1.53467524e+00 -3.57534766e-01 3.06142271e-01
-3.68570447e-01 -1.26604557e+00 -4.90627706e-01 -7.50069857e-01
3.08160335e-01 2.53466904e-01 1.40817249e+00 -1.86643247e-02
-1.37025699e-01 2.88668126e-01 -9.24067497e-01 -3.25370878e-01
-7.65077472e-01 5.39133362e-02 -2.29757786e-01 -1.31306082e-01
6.52643025e-01 -1.94905207e-01 -2.21428558e-01 3.87583487e-02
-2.72547752e-01 -5.78042455e-02 6.50838017e-01 9.95693624e-01
2.78909534e-01 -6.61816955e-01 7.98923910e-01 -1.32500207e+00
8.38217497e-01 -4.75752115e-01 -5.02703071e-01 4.33219254e-01
-1.58026621e-01 -1.01583868e-01 5.15695572e-01 -8.94246772e-02
-1.73838127e+00 -3.01194414e-02 -1.88118353e-01 1.35814592e-01
-1.70123622e-01 6.74364269e-01 -3.45148176e-01 6.73024535e-01
4.81045276e-01 1.59588200e-03 -3.83905977e-01 -4.66929883e-01
8.29978824e-01 8.51617277e-01 5.54107010e-01 -7.32523620e-01
3.80908310e-01 2.46317282e-01 -6.02964818e-01 -1.00903904e+00
-1.51188195e+00 -4.42204356e-01 -1.06769419e+00 -2.71015525e-01
1.22995460e+00 -1.09685004e+00 -3.72255236e-01 6.91066161e-02
-1.75943577e+00 -2.31664076e-01 -7.85892978e-02 4.89904881e-01
-2.41752714e-01 6.26899540e-01 -1.06854665e+00 -1.07484913e+00
-1.13615163e-01 -8.77139330e-01 1.09419096e+00 1.81575924e-01
-2.83934057e-01 -8.68812740e-01 2.52012700e-01 4.59107190e-01
-1.94394648e-01 1.90959170e-01 6.58550560e-01 -1.05573452e+00
-3.06092113e-01 1.46548390e-01 -3.26365344e-02 2.11714223e-01
3.49569432e-02 2.86585111e-02 -1.20514405e+00 6.78678900e-02
3.31006825e-01 -7.28881001e-01 7.01952577e-01 2.03280985e-01
3.42336982e-01 -1.55273154e-01 -2.71871507e-01 -8.78670141e-02
7.93765068e-01 1.61462024e-01 4.40076858e-01 1.85459450e-01
7.12595284e-01 1.19428849e+00 9.47568953e-01 5.25153100e-01
5.88789105e-01 4.82342184e-01 3.50788720e-02 1.09616891e-01
-2.47549601e-02 -1.81524858e-01 6.48162007e-01 1.57151222e+00
4.89327349e-02 -4.54268664e-01 -7.63898849e-01 6.81441784e-01
-1.95739436e+00 -8.88819337e-01 -5.62233269e-01 1.72216892e+00
1.18446171e+00 5.23358881e-01 -1.10623300e-01 -2.36629158e-01
1.03126144e+00 4.67449367e-01 -3.05488139e-01 -1.64176345e-01
-2.75445879e-01 3.34783792e-02 1.17097065e-01 5.54318905e-01
-1.12766707e+00 1.17883921e+00 6.01218939e+00 5.27441204e-01
-4.69228506e-01 4.27823812e-01 6.83548093e-01 5.13450205e-02
-5.63366786e-02 5.58455475e-02 -1.04841316e+00 4.04298484e-01
1.20725572e+00 -5.52388802e-02 -9.67412516e-02 7.71508276e-01
1.80284977e-01 -2.31540963e-01 -1.16686296e+00 5.16915977e-01
1.32403448e-01 -1.17142749e+00 -4.85453308e-01 -3.32935274e-01
4.39578742e-01 6.66693226e-02 -6.34072483e-01 7.48154521e-01
6.78706467e-01 -7.12647736e-01 4.98203129e-01 -2.00453643e-02
5.40490568e-01 -3.97577673e-01 6.34045720e-01 6.41309619e-01
-1.25004601e+00 2.72155315e-01 -4.99172032e-01 -2.88096339e-01
6.53910995e-01 6.61014974e-01 -5.18548131e-01 9.17258382e-01
5.20217538e-01 6.10062957e-01 -2.02135891e-01 2.41460681e-01
-1.10353220e+00 1.04114234e+00 -1.51222860e-02 -3.18066329e-01
1.69700503e-01 -2.37797618e-01 4.76157218e-01 1.59008420e+00
-1.43668115e-01 6.09042823e-01 2.30628550e-01 7.51086414e-01
-1.96539640e-01 1.67859897e-01 -7.20966697e-01 1.68733552e-01
9.31185246e-01 1.46068013e+00 -7.46015251e-01 -5.13340175e-01
-9.51738119e-01 1.03651285e+00 6.06620610e-01 2.47419819e-01
-7.56307781e-01 -2.80311346e-01 1.74953178e-01 -7.66382217e-01
2.55356282e-01 -1.95506543e-01 -1.15410544e-01 -1.46452665e+00
8.09343904e-02 -7.97827482e-01 5.12393236e-01 -2.47844234e-01
-1.45147586e+00 6.03515863e-01 -6.77914917e-02 -1.14647007e+00
-3.67763728e-01 -4.59024996e-01 -8.74199092e-01 6.78048670e-01
-1.57381737e+00 -9.86773372e-01 1.24078944e-01 2.22230107e-01
1.24670053e+00 1.59943700e-01 8.57967317e-01 1.52781978e-01
-9.47392404e-01 3.53781521e-01 -2.38345519e-01 5.75382173e-01
9.39989030e-01 -1.31039286e+00 7.11821198e-01 8.85815382e-01
2.82674760e-01 4.82527345e-01 5.74575365e-01 -6.03357792e-01
-1.35633552e+00 -9.20125365e-01 1.27315950e+00 -7.78138101e-01
8.39150131e-01 -7.17750311e-01 -1.07124674e+00 8.70700955e-01
6.77538693e-01 -1.20125435e-01 9.03620362e-01 3.73287261e-01
-2.43404940e-01 5.70010543e-01 -6.97199523e-01 6.47486389e-01
1.13514245e+00 -5.39234221e-01 -1.22657633e+00 4.34530556e-01
9.25846219e-01 -6.26832366e-01 -8.53337467e-01 2.33580768e-01
1.83175534e-01 -1.25453293e+00 5.98549426e-01 -6.43426120e-01
6.47484720e-01 9.17717367e-02 -9.45491567e-02 -1.24749684e+00
-1.45422192e-02 -7.27922201e-01 -1.11051528e-02 1.74346745e+00
7.54734933e-01 -2.28288949e-01 3.72051090e-01 7.48629451e-01
-5.20062089e-01 -4.15780067e-01 -1.07101834e+00 -7.46471286e-01
-1.83931682e-02 -1.91839695e-01 1.09152757e-01 1.26715350e+00
6.33712947e-01 1.26907265e+00 -6.29548252e-01 6.17127493e-02
2.69654363e-01 2.03434452e-01 8.20197225e-01 -1.02842879e+00
-4.47242647e-01 3.53517830e-02 4.84064728e-01 -1.30628991e+00
4.87810612e-01 -5.73996603e-01 4.53491867e-01 -1.25280261e+00
1.76914200e-01 4.78572622e-02 -6.13820478e-02 1.95875332e-01
-6.16488039e-01 -2.02288404e-01 1.23065531e-01 3.37442368e-01
-9.26451445e-01 7.86697507e-01 1.11972284e+00 1.18296541e-01
-3.24730754e-01 -2.40877911e-01 -6.50227904e-01 9.98352587e-01
6.92415535e-01 -4.82770622e-01 -1.11266561e-01 -4.75122660e-01
-1.72263518e-01 5.23476779e-01 -3.33551437e-01 -2.40322620e-01
1.57733992e-01 -6.80275485e-02 -1.85884595e-01 -5.95438182e-01
2.85919249e-01 -2.10769817e-01 -6.80710673e-01 2.83373415e-01
-5.63220501e-01 -1.49933934e-01 -1.36767805e-01 6.18919790e-01
-4.25936043e-01 -6.56860232e-01 5.04976988e-01 -4.66982633e-01
-6.63566053e-01 -4.14538175e-01 -5.19538403e-01 6.53495908e-01
1.01857257e+00 3.03152740e-01 -6.42757297e-01 -1.01674803e-01
-6.10898018e-01 4.23796862e-01 -2.04605907e-01 4.58814591e-01
2.75390863e-01 -1.02570760e+00 -9.15909111e-01 -2.41081104e-01
1.20127641e-01 1.93006769e-01 -3.05393361e-04 8.10431063e-01
1.12794794e-01 3.97381991e-01 3.50656360e-01 -3.80904466e-01
-1.28774011e+00 4.64695841e-01 -2.64707923e-01 -4.50046778e-01
-7.13400185e-01 9.55575228e-01 2.62599379e-01 -4.34950233e-01
1.81227654e-01 -2.10133836e-01 -6.26353800e-01 4.46527243e-01
6.82566881e-01 5.22319078e-02 -2.19896615e-01 -5.90699077e-01
-3.39924812e-01 1.40842259e-01 -2.56968141e-01 -2.55837411e-01
1.07274652e+00 -4.77211863e-01 -1.76673025e-01 8.99820924e-01
7.44056404e-01 3.38084012e-01 -1.17838550e+00 -3.26703340e-01
8.35498512e-01 -1.25951365e-01 -4.40876663e-01 -4.91863400e-01
-4.40249383e-01 8.69121134e-01 -2.11072072e-01 4.89227057e-01
5.62679708e-01 3.74149382e-01 6.69428051e-01 5.33490539e-01
3.04848820e-01 -1.31983006e+00 3.13751131e-01 9.65270638e-01
7.29526401e-01 -1.41104102e+00 -2.47666687e-01 -9.18959856e-01
-1.07052398e+00 1.04521668e+00 8.74176025e-01 1.94321908e-02
2.22500309e-01 3.20472538e-01 3.18562835e-01 -3.07582110e-01
-1.08180094e+00 -4.09244627e-01 -2.02854797e-01 5.36596596e-01
9.61987615e-01 -1.25578195e-01 -7.16431379e-01 1.09744501e+00
1.01667717e-01 -3.74799281e-01 7.27177799e-01 1.07726669e+00
-5.12162685e-01 -1.22491968e+00 -2.29631379e-01 2.86631584e-02
-6.63762033e-01 -3.38098675e-01 -6.28128946e-01 8.57739449e-01
-4.56195235e-01 1.36495805e+00 -3.36989462e-01 -1.09451987e-01
5.90984941e-01 3.76808912e-01 2.14064315e-01 -1.14606237e+00
-6.67329252e-01 4.21390474e-01 9.15849447e-01 -3.18685293e-01
-8.99405956e-01 -5.74623466e-01 -1.39355135e+00 3.50125402e-01
-6.42380238e-01 2.72156447e-01 2.57087767e-01 1.28853345e+00
1.90421462e-01 6.55829489e-01 8.05588126e-01 -8.06446791e-01
-6.96837902e-01 -1.26758695e+00 -2.53987163e-01 2.79036015e-01
-3.18723246e-02 -4.64948535e-01 -3.18358660e-01 2.22962394e-01]
|
[10.671801567077637, 9.38416862487793]
|
1dafb61e-b344-4691-ac33-1609103c266d
|
distributionally-robust-learning-for-1
|
2010.05784
| null |
https://arxiv.org/abs/2010.05784v3
|
https://arxiv.org/pdf/2010.05784v3.pdf
|
Distributionally Robust Learning for Uncertainty Calibration under Domain Shift
|
We propose a framework for learning calibrated uncertainties under domain shifts. We consider the case where the source (training) distribution differs from the target (test) distribution. We detect such domain shifts through the use of a binary domain classifier and integrate it with the task network and train them jointly end-to-end. The binary domain classifier yields a density ratio that reflects the closeness of a target (test) sample to the source (training) distribution. We employ it to adjust the uncertainty of prediction in the task network. This idea of using the density ratio is based on the distributionally robust learning (DRL) framework, which accounts for the domain shift through adversarial risk minimization. We demonstrate that our method generates calibrated uncertainties that benefit many downstream tasks, such as unsupervised domain adaptation (UDA) and semi-supervised learning (SSL). In these tasks, methods like self-training and FixMatch use uncertainties to select confident pseudo-labels for re-training. Our experiments show that the introduction of DRL leads to significant improvements in cross-domain performance. We also demonstrate that the estimated density ratios show agreement with the human selection frequencies, suggesting a positive correlation with a proxy of human perceived uncertainties.
|
['Anima Anandkumar', 'Junchi Yan', 'Yisong Yue', 'Zhiding Yu', 'Anqi Liu', 'Haoxuan Wang']
|
2020-10-08
|
distributionally-robust-learning-for
|
https://openreview.net/forum?id=FZyZiRYbdK8
|
https://openreview.net/pdf?id=FZyZiRYbdK8
| null |
['density-ratio-estimation']
|
['methodology']
|
[ 4.64489609e-01 4.61038470e-01 -2.61021465e-01 -8.43654692e-01
-1.30599010e+00 -1.04474366e+00 8.15629482e-01 -8.92147608e-03
-6.18859947e-01 1.09490418e+00 6.05089888e-02 -1.59620821e-01
-1.75098836e-01 -5.36390543e-01 -1.13279915e+00 -6.08921766e-01
2.76103765e-01 8.00350666e-01 2.30594561e-01 1.58973217e-01
7.09877834e-02 2.98955858e-01 -1.13200665e+00 2.26626560e-01
1.17310023e+00 1.11202168e+00 -1.86614558e-01 3.82462800e-01
-1.30755724e-02 3.90667260e-01 -9.26496387e-01 -6.24746382e-01
4.55975682e-01 -1.67161033e-01 -6.38704002e-01 -2.44362727e-01
4.81728882e-01 -8.64919275e-02 2.54072342e-02 1.33949935e+00
5.69654107e-01 3.68151754e-01 1.26417339e+00 -1.35887802e+00
-6.34101987e-01 9.38427627e-01 -4.29363608e-01 8.73895213e-02
-4.15841416e-02 7.37179071e-02 6.31639004e-01 -8.86794090e-01
5.95042050e-01 1.42440736e+00 8.81367564e-01 6.74834132e-01
-1.44005907e+00 -1.03968072e+00 3.37528102e-02 -2.63323128e-01
-1.22961104e+00 -5.39886832e-01 6.32389307e-01 -6.26880169e-01
4.47612524e-01 -3.13963830e-01 -2.58164585e-01 1.65380538e+00
1.85357556e-01 4.67433721e-01 1.11975276e+00 -6.33396626e-01
6.96263134e-01 6.55826151e-01 -8.15535635e-02 7.49159083e-02
1.03706881e-01 6.91181898e-01 -5.24984062e-01 -2.82728910e-01
4.96129513e-01 -4.13429290e-01 -5.94199337e-02 -5.92241347e-01
-9.88787830e-01 7.72842586e-01 3.76765788e-01 7.44174570e-02
-3.50904018e-02 1.74818218e-01 4.76997107e-01 3.32356781e-01
8.35259736e-01 5.67650139e-01 -7.06802785e-01 1.51664063e-01
-8.79266739e-01 2.16725513e-01 6.36868119e-01 1.02497196e+00
5.59382141e-01 -3.18595325e-03 -4.99239534e-01 1.14014828e+00
3.25153589e-01 6.55315518e-01 4.50400203e-01 -1.18290758e+00
5.93538105e-01 5.69181256e-02 2.64063329e-01 -4.82795954e-01
-1.90298289e-01 -5.26664436e-01 -5.68769932e-01 4.45930302e-01
9.43002105e-01 -4.39093381e-01 -1.00212908e+00 2.18001485e+00
3.52747828e-01 2.52107501e-01 2.63935506e-01 5.39079070e-01
1.20202459e-01 2.53044754e-01 5.00746310e-01 9.42339152e-02
9.36748087e-01 -5.12121856e-01 -3.61159444e-01 -5.18254638e-01
3.74214768e-01 -4.93328393e-01 1.10575986e+00 3.61799002e-01
-8.26934159e-01 -7.63578057e-01 -1.07702661e+00 4.42692637e-01
-4.44917500e-01 -9.09002796e-02 -9.56599712e-02 7.91709721e-01
-6.91190898e-01 9.48474169e-01 -3.92309129e-01 -1.47386178e-01
4.94680494e-01 2.40882903e-01 -1.78856716e-01 6.36790274e-03
-1.56225383e+00 9.70954657e-01 7.03512788e-01 -5.28076470e-01
-8.91869187e-01 -9.76049483e-01 -9.56317902e-01 -2.02132598e-01
3.10629219e-01 -5.15981019e-01 1.47973216e+00 -1.43747818e+00
-1.43828213e+00 1.01799965e+00 2.07387790e-01 -7.64834344e-01
8.89629245e-01 -1.31349683e-01 -4.33233052e-01 -2.24427119e-01
2.86442220e-01 8.74242961e-01 1.35556161e+00 -1.37571335e+00
-7.22307861e-01 -3.05470794e-01 -4.41486865e-01 1.71423614e-01
-2.17922200e-02 -2.27455527e-01 -1.26801040e-02 -8.79554033e-01
-2.54278958e-01 -9.53374684e-01 -7.74007067e-02 -2.47164369e-02
-5.41475534e-01 -1.59459412e-01 4.41092253e-01 -4.58458871e-01
8.46701920e-01 -2.32144856e+00 -2.03505278e-01 6.80901051e-01
-1.23261325e-02 2.11508825e-01 -3.06323245e-02 -9.17231366e-02
-2.72067070e-01 2.10666731e-01 -6.74885452e-01 -3.90184850e-01
2.56492823e-01 8.73247236e-02 -7.26461947e-01 5.53095222e-01
3.97890329e-01 5.17921329e-01 -8.78881693e-01 -3.14521521e-01
-1.31169604e-02 2.05133483e-01 -4.40019041e-01 3.46566856e-01
-4.52986419e-01 6.77646339e-01 -1.67377964e-01 2.64247209e-01
8.79174054e-01 -3.28696892e-02 8.28342438e-02 -6.90128952e-02
4.07188624e-01 1.73215538e-01 -1.09250617e+00 1.37849832e+00
-4.14015383e-01 3.22964340e-01 -1.15896851e-01 -7.22518742e-01
1.12872362e+00 -4.10030186e-02 -4.34724689e-02 -3.62445354e-01
1.98084004e-02 2.74891675e-01 -1.28004864e-01 9.24025327e-02
1.57471001e-01 -3.73102129e-01 -4.43148553e-01 3.12656790e-01
3.97299320e-01 -2.71682680e-01 -3.06005269e-01 3.55293304e-02
8.32972765e-01 4.26946551e-01 4.92396861e-01 -3.07865500e-01
2.97121346e-01 -1.12252735e-01 5.41032553e-01 9.06029344e-01
-5.20327151e-01 6.53956771e-01 6.58578575e-01 2.33808935e-01
-1.18760300e+00 -1.69530702e+00 -4.93252426e-01 1.36620069e+00
-7.25077093e-02 4.92754847e-01 -8.81515503e-01 -1.22140253e+00
6.72220886e-01 1.27619278e+00 -7.25271106e-01 -5.66569686e-01
-2.00703815e-01 -2.32119203e-01 8.52476418e-01 7.24556267e-01
3.15965623e-01 -8.45970392e-01 -3.74233052e-02 3.23064625e-02
1.13616824e-01 -1.03383565e+00 -6.93480432e-01 6.47723913e-01
-6.68156147e-01 -8.44645560e-01 -7.83880055e-01 -5.25811911e-01
6.28972173e-01 -5.22088230e-01 1.24990439e+00 -7.44322240e-01
3.21472168e-01 5.01761496e-01 -1.18711852e-01 -5.54847479e-01
-1.03600085e+00 -4.20006812e-02 4.49006408e-01 -2.43889377e-01
3.97906989e-01 -4.87742692e-01 -2.62951612e-01 6.14210188e-01
-8.74170244e-01 -5.55113196e-01 3.29214156e-01 9.15466070e-01
6.40628874e-01 -1.34430605e-03 1.14618206e+00 -1.39761901e+00
8.47429454e-01 -6.80464566e-01 -8.76379013e-01 3.36891532e-01
-1.02009821e+00 3.80813271e-01 6.35255516e-01 -7.49480188e-01
-1.42437339e+00 1.85965806e-01 2.04026654e-01 -7.04453647e-01
-4.05018806e-01 1.81895390e-01 -2.60842621e-01 1.56689063e-01
1.36160910e+00 -3.21006387e-01 1.47596709e-02 -1.62687227e-01
6.19105101e-01 8.44439089e-01 8.73559535e-01 -8.93511057e-01
1.00511658e+00 1.21944174e-01 -2.05190092e-01 -1.03778906e-01
-1.09597278e+00 -1.31248742e-01 -8.38353515e-01 -1.61706254e-01
5.12985408e-01 -1.10178709e+00 -2.00798705e-01 3.39420408e-01
-1.03250444e+00 -6.02905691e-01 -6.41162395e-01 3.99565428e-01
-6.61301672e-01 1.59599781e-01 -1.63647398e-01 -8.90511215e-01
4.30351403e-03 -9.27144229e-01 1.03493428e+00 1.20721385e-01
-3.79157841e-01 -1.35689795e+00 1.16625980e-01 1.27955586e-01
3.41767579e-01 2.98834145e-01 9.18302357e-01 -1.38343740e+00
2.91626900e-02 -6.29006848e-02 -1.79355413e-01 8.01043451e-01
-9.33968648e-02 -2.09122255e-01 -1.48335087e+00 -1.47691101e-01
-1.33409828e-01 -4.92663920e-01 8.80257964e-01 5.84532619e-01
1.21915400e+00 6.77329153e-02 -2.13398591e-01 4.76420641e-01
1.08865166e+00 9.52456743e-02 4.27260637e-01 2.10353836e-01
2.86639720e-01 8.43700230e-01 8.59901607e-01 4.10915852e-01
-5.52578904e-02 4.18687046e-01 1.67676225e-01 2.06023186e-01
-1.42424956e-01 -6.44065320e-01 6.03561461e-01 -7.73814544e-02
6.59985900e-01 -1.76885366e-01 -1.10047019e+00 4.63164836e-01
-1.63518643e+00 -7.55676091e-01 3.73608261e-01 2.57224154e+00
1.40688574e+00 5.09012461e-01 1.40776902e-01 -2.88588047e-01
1.10783303e+00 -1.44050658e-01 -1.17501187e+00 -2.21355945e-01
1.60429701e-02 1.78174615e-01 9.03619349e-01 6.54378235e-01
-1.28606474e+00 9.02669966e-01 6.56141663e+00 1.05406153e+00
-7.16292560e-01 3.26772407e-02 9.00018811e-01 1.61304682e-01
-4.09771651e-01 -3.77490461e-01 -8.00638855e-01 6.58979237e-01
1.16435671e+00 -3.81027937e-01 4.03566033e-01 1.16660464e+00
-1.76888734e-01 -4.45077159e-02 -1.51146257e+00 6.85943246e-01
-1.73699006e-01 -8.80128801e-01 -3.66841823e-01 -1.81523293e-01
8.71778429e-01 -5.23122959e-03 4.77258503e-01 6.63919449e-01
9.80157197e-01 -1.01910305e+00 9.16583240e-01 4.61401939e-01
1.33836782e+00 -9.02417243e-01 6.57902539e-01 3.75531793e-01
-5.73947489e-01 -1.38251260e-01 -5.73218465e-01 5.96928716e-01
-1.29453868e-01 1.08025134e+00 -1.29849422e+00 4.03529480e-02
5.39135456e-01 5.13110518e-01 -4.01212215e-01 6.28065944e-01
-3.54457378e-01 6.86219811e-01 -1.77148417e-01 4.80486721e-01
-2.51016319e-01 -2.33758334e-02 6.31896198e-01 1.24545419e+00
3.05811882e-01 -5.36740661e-01 -1.34747565e-01 1.33608818e+00
-4.96904165e-01 -4.46030706e-01 -7.48395562e-01 4.31115665e-02
9.44629848e-01 7.22815752e-01 -3.24499696e-01 -2.30480373e-01
1.91813171e-01 9.87012208e-01 3.57376367e-01 4.54891831e-01
-9.41876173e-01 -3.82632554e-01 7.55515516e-01 -2.47907773e-01
9.94648114e-02 2.67495126e-01 -5.99214017e-01 -7.64119089e-01
-1.63506180e-01 -8.51662338e-01 5.13624191e-01 -7.07758427e-01
-2.16945028e+00 3.98416013e-01 3.41168642e-01 -1.38478851e+00
-6.10627234e-01 -7.09223270e-01 -4.88194019e-01 1.18590474e+00
-1.47207832e+00 -7.60928571e-01 -6.82703592e-03 6.40579522e-01
2.95510739e-01 -4.27016586e-01 6.43953919e-01 7.10955588e-03
-2.69074440e-01 1.07745314e+00 3.67554188e-01 1.11129560e-01
1.64192355e+00 -1.65922701e+00 4.68598217e-01 6.76678002e-01
-2.88991541e-01 4.47638303e-01 8.04477513e-01 -9.05875146e-01
-3.03157240e-01 -1.50313163e+00 5.14061511e-01 -8.53940785e-01
5.96713841e-01 -3.19003284e-01 -9.82199728e-01 6.96021676e-01
-1.15738697e-01 8.85132179e-02 7.91991293e-01 1.53294489e-01
-8.55503261e-01 -1.21517733e-01 -1.85895121e+00 3.14189583e-01
8.87407959e-01 -7.09410727e-01 -7.61726260e-01 3.23702276e-01
1.09233713e+00 -5.27509987e-01 -8.87295723e-01 3.49298447e-01
3.21816087e-01 -8.54063749e-01 9.48893130e-01 -7.04991817e-01
3.19344401e-01 -1.69138029e-01 -3.02255571e-01 -1.95948315e+00
-1.78498656e-01 -2.28681013e-01 1.26675338e-01 1.55366325e+00
7.07009435e-01 -9.51053798e-01 6.02286279e-01 9.47786689e-01
-4.13876921e-02 8.93697236e-03 -1.01797223e+00 -1.04849672e+00
6.42830551e-01 -5.53576708e-01 5.51726282e-01 9.62039113e-01
-1.91604450e-01 4.42482419e-02 -4.61406149e-02 3.61246675e-01
9.35029984e-01 -3.53668213e-01 4.42043692e-01 -1.42942584e+00
-4.19712216e-01 -2.10376069e-01 2.45459899e-02 -6.89361930e-01
7.28431463e-01 -1.03461862e+00 4.45474714e-01 -7.07008421e-01
-1.82622746e-01 -7.32856810e-01 -4.40650433e-01 2.86682785e-01
-1.87037647e-01 -2.13828653e-01 -3.57283093e-03 1.10053867e-01
-4.70127612e-01 4.56017941e-01 8.87910426e-01 4.98966733e-03
-2.04541206e-01 3.41523588e-01 -9.45927501e-01 7.34810054e-01
8.60486150e-01 -8.69298160e-01 -4.05242980e-01 -1.68572694e-01
-5.33565134e-02 -1.28222212e-01 3.02054763e-01 -1.06590748e+00
8.46024975e-02 -2.24283040e-01 8.07025969e-01 -9.08888727e-02
7.59011731e-02 -9.50335324e-01 -2.44853765e-01 9.91714373e-02
-9.49411988e-01 -4.87807244e-01 3.83798033e-01 1.03082120e+00
8.16662516e-03 -4.25356716e-01 1.25198555e+00 1.86669141e-01
-5.00587046e-01 -7.23323524e-02 -1.55853992e-02 5.65710008e-01
1.04533184e+00 8.94221067e-02 -4.52664733e-01 -4.15099651e-01
-7.74968505e-01 2.71095335e-01 4.55600679e-01 2.03526840e-01
3.26423109e-01 -1.31595850e+00 -6.32217586e-01 2.54975736e-01
3.16358149e-01 2.19660223e-01 -1.53816538e-02 2.53553957e-01
2.62997955e-01 3.52760516e-02 -7.93262348e-02 -6.74402118e-01
-6.99416161e-01 4.11475986e-01 6.27156258e-01 -2.11082682e-01
7.89700598e-02 1.08409643e+00 2.55578518e-01 -9.77229834e-01
5.49529374e-01 -2.17603117e-01 1.44721374e-01 3.44294094e-04
3.43400061e-01 2.96654552e-01 -1.18878424e-01 -1.82004511e-01
-3.22925985e-01 2.53214210e-01 1.12179652e-01 -4.09292161e-01
9.01976705e-01 -8.72688890e-02 4.51756299e-01 5.10722160e-01
1.00665033e+00 1.39369562e-01 -1.74560392e+00 -6.24525785e-01
3.47516179e-01 -4.65324193e-01 -1.55167624e-01 -1.42826295e+00
-5.60582280e-01 7.45531917e-01 7.88785577e-01 -9.75651070e-02
9.20087755e-01 1.38853282e-01 2.88090706e-01 2.20047504e-01
3.18592280e-01 -1.37916219e+00 1.09204985e-01 5.67846000e-01
9.29392755e-01 -1.41177166e+00 -2.74028510e-01 -2.02421114e-01
-1.27044272e+00 8.79123092e-01 6.69050753e-01 -2.16366142e-01
7.56282866e-01 3.97202373e-01 1.94262251e-01 4.67427582e-01
-5.23692250e-01 -1.90443266e-03 4.89532858e-01 1.30819917e+00
1.66576073e-01 1.89355537e-01 3.75150204e-01 9.32495952e-01
-1.22682929e-01 -7.59262890e-02 1.68447331e-01 3.66087854e-01
-5.09959102e-01 -1.12566638e+00 -5.89311540e-01 4.61226732e-01
-3.12937260e-01 3.46404165e-02 -3.95489484e-01 5.51837146e-01
1.47171274e-01 9.09251332e-01 1.55117169e-01 -3.36160570e-01
4.86172974e-01 6.42269194e-01 2.73583829e-01 -7.15277791e-01
-4.71210063e-01 -1.77330792e-01 1.07978024e-01 -2.87012488e-01
-1.12546727e-01 -5.93174458e-01 -1.29646587e+00 -1.80210471e-02
-1.41305715e-01 -5.79983033e-02 5.93427002e-01 8.15915287e-01
3.27168584e-01 4.50052977e-01 6.11188114e-01 -6.97589278e-01
-1.45171785e+00 -1.00114810e+00 -8.27035844e-01 6.11264110e-01
2.63331532e-01 -9.57799077e-01 -7.86971867e-01 5.49744256e-02]
|
[10.245031356811523, 3.2294323444366455]
|
a741a564-7b90-426e-a990-ad427828df0d
|
composite-learning-for-robust-and-effective
|
2210.07239
| null |
https://arxiv.org/abs/2210.07239v1
|
https://arxiv.org/pdf/2210.07239v1.pdf
|
Composite Learning for Robust and Effective Dense Predictions
|
Multi-task learning promises better model generalization on a target task by jointly optimizing it with an auxiliary task. However, the current practice requires additional labeling efforts for the auxiliary task, while not guaranteeing better model performance. In this paper, we find that jointly training a dense prediction (target) task with a self-supervised (auxiliary) task can consistently improve the performance of the target task, while eliminating the need for labeling auxiliary tasks. We refer to this joint training as Composite Learning (CompL). Experiments of CompL on monocular depth estimation, semantic segmentation, and boundary detection show consistent performance improvements in fully and partially labeled datasets. Further analysis on depth estimation reveals that joint training with self-supervision outperforms most labeled auxiliary tasks. We also find that CompL can improve model robustness when the models are evaluated in new domains. These results demonstrate the benefits of self-supervision as an auxiliary task, and establish the design of novel task-specific self-supervised methods as a new axis of investigation for future multi-task learning research.
|
['Luc van Gool', 'Fisher Yu', 'David Bruggemann', 'Thomas E. Huang', 'Menelaos Kanakis']
|
2022-10-13
| null | null | null | null |
['boundary-detection']
|
['computer-vision']
|
[ 4.30840343e-01 2.56756097e-01 -5.62653482e-01 -8.03346038e-01
-1.26575971e+00 -5.03916383e-01 5.11001706e-01 -1.56464159e-01
-4.83854979e-01 5.86095214e-01 2.02885002e-01 -8.95161703e-02
3.37475747e-01 -3.10086012e-01 -8.05382311e-01 -7.06972957e-01
3.92996043e-01 6.54160142e-01 3.11921746e-01 5.03831089e-01
3.13375075e-03 2.18003616e-01 -1.33573508e+00 4.87191916e-01
1.01145065e+00 8.83643508e-01 6.62908137e-01 4.14951622e-01
2.01537415e-01 6.14023030e-01 -5.08815646e-01 -1.44675180e-01
3.17413539e-01 -3.28604579e-02 -9.37174022e-01 5.66614091e-01
9.46623564e-01 -3.90958458e-01 1.61220029e-01 7.30037749e-01
4.91940409e-01 1.38237774e-01 6.74106359e-01 -1.24921632e+00
-3.32816899e-01 9.89430472e-02 -6.74812078e-01 8.09106454e-02
-1.01145357e-01 4.60110279e-03 1.09758723e+00 -7.18302906e-01
6.45290256e-01 1.15274417e+00 7.85289288e-01 6.27231658e-01
-1.32069325e+00 -6.46528125e-01 4.50316787e-01 -8.99753496e-02
-9.99252200e-01 -5.32094657e-01 6.62660718e-01 -5.17370880e-01
9.01084304e-01 -3.26661795e-01 1.42252028e-01 1.11996174e+00
-3.12907808e-02 1.41145241e+00 1.13572073e+00 -1.73877090e-01
1.17298208e-01 3.99031848e-01 2.36250892e-01 7.82213688e-01
3.33626121e-01 1.64521858e-01 -4.73618448e-01 9.85147208e-02
7.69068301e-01 -3.04852933e-01 -1.04226954e-01 -6.02572560e-01
-9.82376814e-01 6.98451340e-01 2.22071454e-01 1.92886978e-01
-1.62947163e-01 1.56390250e-01 4.08242047e-01 1.63454667e-01
1.12179601e+00 3.02643389e-01 -8.74970734e-01 3.52994353e-02
-1.10565794e+00 1.88723743e-01 3.61736625e-01 1.15504336e+00
1.18678987e+00 9.45143998e-02 -1.51329832e-02 1.03521359e+00
1.93650693e-01 4.39442903e-01 3.39256436e-01 -1.19770789e+00
5.49305499e-01 5.91150522e-01 -4.57716696e-02 -2.24965304e-01
-8.05502355e-01 -8.31742823e-01 -4.36141193e-01 1.23132743e-01
6.19702280e-01 -4.25642490e-01 -9.68684793e-01 1.90286481e+00
4.40944076e-01 1.31694049e-01 4.53264862e-02 6.37582004e-01
6.84020877e-01 3.56411546e-01 2.94984698e-01 -3.80664393e-02
9.73850667e-01 -1.52853322e+00 -4.95006382e-01 -9.30148602e-01
1.27905250e+00 -5.66503108e-01 1.10310054e+00 4.62916613e-01
-9.35997069e-01 -7.72069812e-01 -8.84276450e-01 -1.82394326e-01
-5.73972650e-02 4.66797620e-01 8.89615119e-01 4.56582546e-01
-1.03596365e+00 4.28686708e-01 -9.32675362e-01 -2.21118003e-01
6.86529219e-01 2.68062383e-01 -4.81801957e-01 -3.27931166e-01
-5.70739210e-01 9.94813561e-01 2.80045450e-01 -3.14118624e-01
-1.11035717e+00 -8.22506368e-01 -1.25678265e+00 -2.84716219e-01
3.65006447e-01 -6.80860817e-01 1.51911116e+00 -1.08533061e+00
-1.01495075e+00 1.35270917e+00 -5.01773596e-01 -4.36473966e-01
4.53018248e-01 -5.07303596e-01 9.00163800e-02 1.71622783e-01
7.08029151e-01 1.31210375e+00 1.06935620e+00 -1.49441707e+00
-9.79597270e-01 -7.01885283e-01 -2.79352460e-02 5.31462431e-01
-3.06108236e-01 -3.31034601e-01 -5.32717407e-01 -5.74400842e-01
2.72143275e-01 -1.09004104e+00 -1.68666691e-01 -4.10164110e-02
-3.01484793e-01 -4.27261293e-01 1.07165515e+00 -5.66946685e-01
8.23318481e-01 -2.24249673e+00 2.99012125e-01 -3.99977207e-01
1.27613991e-01 1.20499380e-01 -2.43258268e-01 -1.96757428e-02
-1.46225452e-01 -3.44944894e-02 -3.31002623e-01 -1.16450047e+00
-3.83069754e-01 3.90100569e-01 1.58718720e-01 5.19871294e-01
1.78608447e-01 1.00665605e+00 -7.25903094e-01 -6.42827213e-01
2.37417206e-01 -4.33397889e-02 -5.86863697e-01 1.93933845e-01
-4.68256682e-01 8.27084601e-01 -2.36246049e-01 7.52438426e-01
5.22538126e-01 -6.25617027e-01 -8.52625743e-02 -2.94047564e-01
1.54265374e-01 3.46508831e-01 -9.29199457e-01 1.94010675e+00
-6.51367903e-01 7.02822149e-01 4.05429870e-01 -1.18418169e+00
7.45680749e-01 2.20686644e-01 4.95556802e-01 -6.50753260e-01
-1.50639892e-01 1.60578638e-01 -7.63216242e-02 -5.22078753e-01
2.00447753e-01 -2.99601912e-01 1.54875845e-01 5.96142352e-01
2.93428600e-01 -3.37448567e-01 7.86833540e-02 8.85576569e-03
7.77076185e-01 4.58692759e-01 4.14465480e-02 -2.10068345e-01
1.81525573e-01 1.89087808e-01 7.55740881e-01 5.52283704e-01
-5.36147475e-01 5.51077843e-01 6.14364408e-02 -2.90019691e-01
-9.80774641e-01 -7.45200813e-01 -3.55413347e-01 1.39703262e+00
9.04041231e-02 -7.32763931e-02 -6.45302951e-01 -1.23297322e+00
2.77859896e-01 6.98436260e-01 -5.56855798e-01 6.72197342e-02
-4.21246141e-01 -8.56212795e-01 3.78006160e-01 9.64974046e-01
7.05980122e-01 -8.01518857e-01 -5.05241156e-01 4.97522280e-02
-2.77342856e-01 -1.63764048e+00 -3.81871045e-01 6.14860773e-01
-1.32016790e+00 -1.03510225e+00 -1.00895452e+00 -1.04059386e+00
7.18956351e-01 5.79771996e-01 1.00623524e+00 -1.79708883e-01
1.44744292e-01 5.82855046e-01 -2.17842326e-01 -3.16030741e-01
-2.29606256e-01 4.16434139e-01 -1.78897023e-01 -1.71511680e-01
2.40722433e-01 -4.76003170e-01 -3.78528625e-01 4.22865719e-01
-6.74230993e-01 4.90208596e-01 5.77003360e-01 8.52132022e-01
5.80374420e-01 -2.27429181e-01 9.02163029e-01 -1.09980059e+00
1.15071177e-01 -2.75795251e-01 -4.27552462e-01 1.04788125e-01
-7.71999240e-01 1.56756088e-01 1.92400739e-01 -2.53308177e-01
-1.53950453e+00 3.18275452e-01 1.57228224e-02 -3.63252670e-01
-3.36413622e-01 2.80560523e-01 -9.37296599e-02 -9.03222933e-02
6.90019906e-01 3.82175148e-02 -1.38918729e-02 -6.72496736e-01
1.12960182e-01 5.20249307e-01 2.56915420e-01 -5.44686496e-01
5.43729007e-01 7.43882120e-01 2.02552206e-03 -7.80503333e-01
-1.60774243e+00 -7.73324192e-01 -9.11708176e-01 4.41943668e-02
1.03442144e+00 -1.45919335e+00 -8.06808472e-02 6.58109546e-01
-1.08915949e+00 -1.01748872e+00 -6.33583441e-02 4.98794556e-01
-6.96388543e-01 5.79146266e-01 -4.71530437e-01 -6.64385080e-01
-3.85369500e-03 -1.01651847e+00 1.70285869e+00 -5.40146641e-02
-2.05084622e-01 -1.52407515e+00 -5.48317237e-03 1.11667955e+00
-6.47720397e-02 -1.85150176e-01 8.31261814e-01 -6.73982620e-01
-6.59989834e-01 -4.06134278e-02 -4.24693406e-01 7.58273244e-01
3.08820933e-01 -4.79624122e-01 -1.28630066e+00 -4.11557972e-01
1.16286188e-01 -7.97300041e-01 1.23991883e+00 7.19373882e-01
1.01285112e+00 2.71357924e-01 -4.96524215e-01 8.99675369e-01
1.23153830e+00 4.32340465e-02 1.39256030e-01 3.90188873e-01
9.65273976e-01 7.99082637e-01 9.36371386e-01 1.28341585e-01
6.44796729e-01 5.53265750e-01 3.26419234e-01 -4.50145990e-01
-3.82377326e-01 -1.27379656e-01 3.36024374e-01 1.33434311e-01
3.29434961e-01 1.62804779e-02 -1.05253124e+00 6.44200027e-01
-1.81103933e+00 -5.59528232e-01 2.11906433e-02 1.88216305e+00
8.28367949e-01 2.29142100e-01 1.26071960e-01 8.57959094e-04
5.31746149e-01 2.36556008e-01 -9.70581293e-01 7.71858171e-02
-2.53602505e-01 -2.32730657e-02 6.18204117e-01 6.40460730e-01
-1.54596543e+00 1.40405595e+00 6.52549267e+00 6.70506001e-01
-1.00625110e+00 3.88788015e-01 8.71742249e-01 3.50031815e-02
-1.88401714e-01 -6.08829185e-02 -1.27037776e+00 4.33964580e-02
4.87375081e-01 1.20220840e-01 -9.75461155e-02 1.19167292e+00
2.46433556e-01 -6.05052173e-01 -1.49905717e+00 1.03672707e+00
2.49679118e-01 -1.00011146e+00 -1.76077783e-01 5.21917157e-02
1.20776415e+00 2.69091785e-01 5.82141317e-02 2.93474674e-01
2.27920502e-01 -9.36872005e-01 4.37514544e-01 -1.03818603e-01
6.84461415e-01 -2.14356139e-01 6.67857885e-01 6.77174270e-01
-9.29292679e-01 -1.11545913e-01 -1.57138109e-01 -2.14272290e-01
3.09663773e-01 6.08058035e-01 -7.80140400e-01 3.30798626e-01
5.19926727e-01 1.21747148e+00 -6.69556856e-01 8.30869973e-01
-3.53060663e-01 5.29184818e-01 -2.62131453e-01 4.72438008e-01
3.29478294e-01 -3.57244685e-02 3.43408942e-01 1.15897942e+00
-2.76533663e-01 -3.03423285e-01 4.99438047e-01 6.51978493e-01
1.55198732e-02 -4.80300263e-02 -6.22717857e-01 1.05411798e-01
2.44026259e-01 1.04793632e+00 -6.58492625e-01 -4.92153198e-01
-6.49826944e-01 1.05369139e+00 5.34544587e-01 4.76123005e-01
-5.05826354e-01 3.86398733e-01 5.99818349e-01 -7.50548169e-02
1.75506920e-01 -3.81825656e-01 -1.03048408e+00 -1.20079780e+00
6.59519807e-02 -6.66464925e-01 3.57996076e-01 -7.73673892e-01
-1.17644143e+00 1.23161629e-01 1.67799532e-01 -8.83758664e-01
-1.87778503e-01 -7.14997053e-01 -3.24093938e-01 7.00742424e-01
-1.78789449e+00 -1.42875922e+00 -4.67134982e-01 5.32240868e-01
9.78175282e-01 -5.70602790e-02 5.60994565e-01 3.13307762e-01
-7.53963828e-01 5.64207554e-01 -1.05670765e-01 -1.16416410e-01
1.02836537e+00 -1.34706807e+00 3.06126267e-01 7.02833772e-01
2.48273075e-01 -5.18658459e-02 4.04239774e-01 -8.66454780e-01
-7.27702081e-01 -1.26330090e+00 7.00468183e-01 -5.16242921e-01
2.81246126e-01 -2.20674410e-01 -8.75710249e-01 1.32591653e+00
-9.66590941e-02 -2.46809587e-01 5.90632081e-01 3.36245269e-01
-2.09555522e-01 -9.38678812e-03 -9.35836375e-01 2.72520572e-01
1.21142483e+00 -6.07887805e-01 -5.08994162e-01 7.29045391e-01
8.43120098e-01 -3.57688963e-01 -7.29706705e-01 7.32472241e-01
2.34444126e-01 -1.03412735e+00 9.38026547e-01 -4.38480854e-01
3.89083833e-01 2.15653211e-01 -1.38545036e-01 -1.32607377e+00
-1.87566742e-01 -1.67246073e-01 4.06208597e-02 1.03059924e+00
5.45608878e-01 -4.86552298e-01 1.37366617e+00 5.82116008e-01
-5.70350528e-01 -7.03358948e-01 -7.04434156e-01 -9.65718031e-01
3.68751377e-01 -6.46225989e-01 -1.58213884e-01 9.15304005e-01
-2.85250008e-01 5.86167097e-01 -2.85646826e-01 1.53233007e-01
8.71330023e-01 3.91329229e-02 9.21757877e-01 -1.26330328e+00
-2.14106470e-01 -1.95146620e-01 -1.93272382e-02 -1.68200421e+00
6.84539795e-01 -9.08284307e-01 2.22107917e-02 -1.77188087e+00
2.52500862e-01 -6.37117326e-01 2.07591370e-01 7.42168844e-01
-4.06777471e-01 1.07333839e-01 -2.54333690e-02 3.53691906e-01
-6.28656030e-01 6.83593333e-01 1.54437447e+00 1.43647240e-02
-1.92741692e-01 3.71054053e-01 -6.54509187e-01 9.33841586e-01
6.91020787e-01 -3.63451988e-01 -6.60428405e-01 -7.14535475e-01
-2.68331766e-01 -9.01260376e-02 2.40031824e-01 -9.59748268e-01
8.98535773e-02 -9.19887945e-02 3.55264604e-01 -6.87573791e-01
5.83985806e-01 -7.01602399e-01 -4.18711275e-01 2.62393445e-01
-3.23683441e-01 -2.74106055e-01 4.40850645e-01 4.91236895e-01
-2.57173955e-01 -3.15405756e-01 1.00077772e+00 -1.50328308e-01
-1.02333295e+00 2.67384261e-01 -7.33921304e-02 4.60565060e-01
9.96772885e-01 -3.84478599e-01 -2.31397346e-01 -4.48174030e-01
-1.05629957e+00 5.01233280e-01 4.96865600e-01 2.35227302e-01
3.34782362e-01 -9.18847859e-01 -6.55678332e-01 1.69726163e-02
9.41648483e-02 5.10186672e-01 1.77729368e-01 7.55537987e-01
3.79726500e-03 6.24495149e-01 -4.97323088e-02 -1.07923102e+00
-1.37659371e+00 2.86118269e-01 3.95836979e-01 -4.12258148e-01
-3.94084126e-01 1.13272560e+00 7.76051879e-01 -5.80996037e-01
4.48672324e-01 -1.75415665e-01 1.50583368e-02 1.67263582e-01
3.41339372e-02 3.83107334e-01 -1.87712312e-02 -3.52810085e-01
-1.54389903e-01 6.56727314e-01 -2.16518581e-01 -1.21139906e-01
1.31211448e+00 -3.22541475e-01 3.47539097e-01 6.24027610e-01
1.26985800e+00 -4.72148776e-01 -1.84825051e+00 -5.00016868e-01
2.32043371e-01 -2.78808951e-01 2.06162184e-01 -8.77407789e-01
-1.03656316e+00 1.05217266e+00 2.08681747e-01 -3.82855028e-01
1.09429932e+00 2.65696526e-01 6.15706503e-01 4.07516837e-01
3.51909012e-01 -1.19441640e+00 4.74075258e-01 4.66387123e-01
5.87849438e-01 -1.88180840e+00 -1.25120908e-01 -7.93531716e-01
-9.30882990e-01 5.80984473e-01 1.18406296e+00 5.93122840e-02
6.74284995e-01 2.12930068e-01 1.43872619e-01 -1.44061193e-01
-7.31929898e-01 -5.01354039e-01 3.72849613e-01 7.55672574e-01
4.03588086e-01 -2.64423043e-01 1.20012581e-01 3.85733217e-01
2.16052420e-02 -6.28314316e-02 1.37529865e-01 9.60597277e-01
-5.63861728e-01 -1.17280495e+00 -1.01511464e-01 4.64594781e-01
-2.24481001e-01 -1.43086165e-01 -3.54149073e-01 8.72226894e-01
1.81841329e-01 9.38650429e-01 2.28814766e-01 -9.20665115e-02
9.38498452e-02 4.50928479e-01 5.72294831e-01 -1.30897021e+00
-4.94169801e-01 2.30369672e-01 3.28690171e-01 -4.81764078e-01
-5.27510166e-01 -8.53302896e-01 -1.07538927e+00 2.36033618e-01
-3.52275908e-01 -1.49907321e-01 6.37899518e-01 1.27624869e+00
3.17121327e-01 3.94843400e-01 3.91338676e-01 -8.45446765e-01
-5.38975537e-01 -1.07222354e+00 -4.08044398e-01 5.16248465e-01
3.61445367e-01 -8.94234896e-01 -5.02949536e-01 1.81442812e-01]
|
[9.524868965148926, 1.4347320795059204]
|
805c8b36-45d7-44ef-a7b1-a9bf9fc09c9d
|
from-recognition-to-cognition-visual
|
1811.10830
| null |
http://arxiv.org/abs/1811.10830v2
|
http://arxiv.org/pdf/1811.10830v2.pdf
|
From Recognition to Cognition: Visual Commonsense Reasoning
|
Visual understanding goes well beyond object recognition. With one glance at
an image, we can effortlessly imagine the world beyond the pixels: for
instance, we can infer people's actions, goals, and mental states. While this
task is easy for humans, it is tremendously difficult for today's vision
systems, requiring higher-order cognition and commonsense reasoning about the
world. We formalize this task as Visual Commonsense Reasoning. Given a
challenging question about an image, a machine must answer correctly and then
provide a rationale justifying its answer.
Next, we introduce a new dataset, VCR, consisting of 290k multiple choice QA
problems derived from 110k movie scenes. The key recipe for generating
non-trivial and high-quality problems at scale is Adversarial Matching, a new
approach to transform rich annotations into multiple choice questions with
minimal bias. Experimental results show that while humans find VCR easy (over
90% accuracy), state-of-the-art vision models struggle (~45%).
To move towards cognition-level understanding, we present a new reasoning
engine, Recognition to Cognition Networks (R2C), that models the necessary
layered inferences for grounding, contextualization, and reasoning. R2C helps
narrow the gap between humans and machines (~65%); still, the challenge is far
from solved, and we provide analysis that suggests avenues for future work.
|
['Yejin Choi', 'Rowan Zellers', 'Ali Farhadi', 'Yonatan Bisk']
|
2018-11-27
|
from-recognition-to-cognition-visual-1
|
http://openaccess.thecvf.com/content_CVPR_2019/html/Zellers_From_Recognition_to_Cognition_Visual_Commonsense_Reasoning_CVPR_2019_paper.html
|
http://openaccess.thecvf.com/content_CVPR_2019/papers/Zellers_From_Recognition_to_Cognition_Visual_Commonsense_Reasoning_CVPR_2019_paper.pdf
|
cvpr-2019-6
|
['multiple-choice-qa', 'visual-commonsense-reasoning']
|
['natural-language-processing', 'reasoning']
|
[ 6.23503685e-01 5.54768503e-01 1.57157630e-01 -5.32748103e-01
-7.29462385e-01 -7.76045263e-01 6.93909645e-01 3.42999399e-03
-3.92556548e-01 6.10321939e-01 3.33824277e-01 -7.35975027e-01
6.17294759e-02 -6.76800132e-01 -7.79247463e-01 -9.35834795e-02
6.54960096e-01 4.42822129e-01 4.82443646e-02 -5.08823633e-01
3.37011963e-01 1.57700479e-01 -1.49178290e+00 8.25427413e-01
7.86862135e-01 8.30883682e-01 1.31480381e-01 7.87014961e-01
5.82827926e-02 1.54267681e+00 -5.53157091e-01 -1.24489617e+00
2.40368128e-01 -6.49984062e-01 -1.46950936e+00 1.19532496e-01
1.01752114e+00 -3.66113394e-01 -3.55134070e-01 1.46007192e+00
-8.66645873e-02 1.16181269e-01 6.06191099e-01 -1.45283449e+00
-1.46133864e+00 3.79467547e-01 -1.20794125e-01 1.51719347e-01
6.28036141e-01 8.11713755e-01 1.22802413e+00 -6.45472527e-01
7.28367686e-01 1.49027300e+00 4.66440201e-01 1.03946233e+00
-1.24956453e+00 -5.13503909e-01 2.36039609e-01 7.70846725e-01
-9.13242042e-01 -5.50947070e-01 5.84176183e-01 -6.81398511e-01
1.00589752e+00 5.28222322e-01 7.09760368e-01 1.23184812e+00
-6.70119897e-02 6.59098089e-01 1.50646162e+00 -2.45309561e-01
1.49406865e-01 -2.72802785e-02 3.59786004e-01 7.38008320e-01
2.41854966e-01 -1.80876162e-02 -4.83274400e-01 1.31604195e-01
6.95716321e-01 -1.51648819e-01 -3.24103266e-01 -8.33165944e-02
-1.53998327e+00 6.75852537e-01 7.44740248e-01 -9.29178111e-03
-2.18088120e-01 3.48809481e-01 9.62978229e-02 3.80658269e-01
-1.72993541e-01 1.00478494e+00 -1.40988350e-01 -8.33901912e-02
-6.49953127e-01 4.87536550e-01 8.13837290e-01 7.77727902e-01
7.33666003e-01 -2.11834013e-01 -3.85459512e-01 2.90385306e-01
3.21694970e-01 5.68729281e-01 2.68884711e-02 -1.80292845e+00
4.85113025e-01 6.51874542e-01 2.89736062e-01 -1.31596017e+00
-7.56092593e-02 -2.50987440e-01 -7.67697036e-01 4.02830392e-01
8.55789900e-01 5.98526336e-02 -9.18130219e-01 1.84793091e+00
-7.83278942e-02 1.55067474e-01 2.70394862e-01 1.31296349e+00
1.02972567e+00 3.81406218e-01 1.95845574e-01 2.51675993e-01
1.80642271e+00 -9.91299391e-01 -5.80622375e-01 -8.47966433e-01
1.75515816e-01 -4.26163167e-01 1.42046225e+00 3.92191559e-01
-1.10717881e+00 -5.45871019e-01 -1.06177235e+00 -6.38032913e-01
-3.00596476e-01 -2.52642989e-01 8.56320620e-01 3.87728959e-01
-1.06693983e+00 3.75506073e-01 -2.71172166e-01 -3.01738739e-01
7.57529318e-01 -1.70751885e-01 -4.27309245e-01 -4.90897387e-01
-1.26784730e+00 1.37814391e+00 2.06111327e-01 9.69052613e-02
-9.98833895e-01 -5.64801455e-01 -1.00709629e+00 7.08656162e-02
6.22042537e-01 -1.07657707e+00 1.41036475e+00 -1.21705127e+00
-9.25321877e-01 1.45379019e+00 -2.82383561e-01 -7.67543554e-01
5.04635215e-01 -2.31192484e-01 -3.30457836e-01 3.65143508e-01
2.21529812e-01 9.73817408e-01 7.88542509e-01 -1.34331882e+00
-3.93363923e-01 -2.30493963e-01 8.85609925e-01 1.37327164e-01
2.64934957e-01 6.55751750e-02 -2.96743304e-01 -3.94417286e-01
1.28491670e-01 -8.03489923e-01 -1.92868605e-01 4.73926008e-01
-1.88178882e-01 -1.85796157e-01 4.66197990e-02 -8.61962736e-01
6.41466379e-01 -1.98840165e+00 6.24361895e-02 -2.20790476e-01
6.76932275e-01 1.57971203e-01 -1.98654473e-01 -4.21263278e-03
-1.09403387e-01 2.48211697e-01 -2.30915487e-01 -1.52319252e-01
3.53479236e-01 3.84918511e-01 -7.56034017e-01 2.06453696e-01
5.21221757e-01 1.39019525e+00 -1.29021406e+00 -4.80770946e-01
2.76862651e-01 6.41341656e-02 -7.14148283e-01 4.09373641e-02
-5.63660443e-01 2.61098593e-01 -2.69878320e-02 5.06477594e-01
4.34093893e-01 -7.69646764e-01 1.27129376e-01 -3.49937648e-01
3.53030890e-01 3.40984166e-01 -9.64613199e-01 1.68064106e+00
-1.09564513e-01 1.06879175e+00 -1.98041141e-01 -8.35729003e-01
6.08330965e-01 -7.14834630e-02 -3.93596500e-01 -8.87552142e-01
-7.10868314e-02 -1.81139007e-01 2.23731041e-01 -7.05787778e-01
4.42626685e-01 -3.80183220e-01 -1.18039884e-01 3.30339789e-01
-1.14098284e-02 -4.90717888e-01 6.71569258e-02 5.01646996e-01
1.15877354e+00 5.92400618e-02 3.60667557e-01 -4.22843127e-03
3.70464444e-01 6.16245151e-01 5.29931843e-01 1.03755784e+00
-5.68934023e-01 5.66695452e-01 4.83391911e-01 -8.31475019e-01
-9.48095083e-01 -1.23266160e+00 3.61245245e-01 9.40922856e-01
3.76192600e-01 -3.86172473e-01 -7.30194926e-01 -4.12066400e-01
-1.09307446e-01 1.09206629e+00 -8.18102121e-01 -1.12642087e-01
-4.41759318e-01 -1.82758152e-01 7.64776230e-01 5.13463140e-01
8.59274626e-01 -1.18238914e+00 -9.22124624e-01 -2.25778520e-01
-7.85948873e-01 -1.47399211e+00 -9.08645093e-02 -3.32786947e-01
-4.48225975e-01 -1.39643764e+00 -1.85973287e-01 -4.34250236e-01
6.75114870e-01 4.44003373e-01 1.76795053e+00 4.00120199e-01
-3.83919835e-01 6.21265411e-01 -1.98891461e-01 -4.10686821e-01
-5.84679306e-01 -7.66068280e-01 -2.68413723e-01 -2.04259768e-01
5.95362067e-01 -3.09024751e-01 -7.02928305e-01 1.67748049e-01
-6.23057246e-01 5.16389608e-01 4.76055086e-01 7.02412307e-01
4.25460458e-01 -1.84521437e-01 2.29818448e-01 -6.54873788e-01
6.17228806e-01 -3.14967662e-01 -2.72328794e-01 5.31173229e-01
-3.36872995e-01 1.01061255e-01 5.36225200e-01 -3.02291989e-01
-1.17162168e+00 -2.67615914e-01 1.15476161e-01 -3.86725783e-01
-4.23596025e-01 1.62144095e-01 -6.80369809e-02 1.53385893e-01
1.03636014e+00 2.03018412e-01 -1.30018815e-01 -6.75208271e-02
9.65179205e-01 2.81480789e-01 1.20737386e+00 -8.36155891e-01
8.51556242e-01 7.86251664e-01 -8.49664509e-02 -4.59230632e-01
-1.46871865e+00 -1.11555099e-01 -4.45595115e-01 -3.19124967e-01
1.34005189e+00 -9.34288025e-01 -1.24136543e+00 5.55951335e-02
-1.48171520e+00 -4.33319420e-01 -3.47671360e-01 -2.69483458e-02
-6.94862604e-01 3.47883224e-01 -4.46996957e-01 -7.03382492e-01
-9.12235603e-02 -9.73452449e-01 6.74713850e-01 2.50948131e-01
-5.90794981e-01 -6.93316698e-01 -2.37645805e-01 1.29019415e+00
2.88875371e-01 2.89914995e-01 1.07141984e+00 -4.77226138e-01
-9.81329381e-01 3.18204731e-01 -7.61603117e-01 4.48189795e-01
-2.93447137e-01 -3.32591057e-01 -9.49813843e-01 8.40645730e-02
2.47313343e-02 -8.64780188e-01 8.31517339e-01 -1.79341316e-01
1.27895033e+00 -3.28581452e-01 1.22694187e-01 3.72233897e-01
1.25485301e+00 -3.40330265e-02 1.09512460e+00 3.34335625e-01
5.16597569e-01 7.06199706e-01 6.12837255e-01 1.09242892e-03
1.01390851e+00 3.31609040e-01 7.94372261e-01 -9.51932520e-02
-4.95987624e-01 -3.10976952e-01 1.33844554e-01 2.24535048e-01
-3.32625329e-01 9.32393968e-02 -1.16657114e+00 6.17655814e-01
-1.86189055e+00 -1.44857323e+00 -2.24307597e-01 1.70872366e+00
9.42473292e-01 2.36332834e-01 -4.10151422e-01 -5.67280501e-02
4.63146955e-01 5.57854585e-02 -7.77412951e-01 -5.18643677e-01
-2.38166198e-01 7.92151466e-02 -2.51944233e-02 5.30081868e-01
-8.51571381e-01 1.07379830e+00 6.70319700e+00 6.04891539e-01
-6.56810224e-01 4.63918671e-02 5.72011948e-01 1.85507298e-01
-6.09878302e-01 4.00747985e-01 -4.16684628e-01 6.05153367e-02
4.70822334e-01 5.72122931e-02 9.51869547e-01 6.86386645e-01
-2.79687464e-01 -1.86927259e-01 -1.44097877e+00 1.39653909e+00
4.81983721e-01 -1.63646352e+00 2.41129234e-01 -2.48717278e-01
5.46032906e-01 -1.00568593e-01 5.09357452e-02 5.14774978e-01
7.42354095e-01 -1.63358963e+00 8.82412970e-01 8.23212922e-01
6.57878578e-01 -2.53231466e-01 3.91219735e-01 4.88955438e-01
-6.41422212e-01 -1.14203095e-01 -4.56140310e-01 -7.70543218e-01
-7.63841942e-02 2.34824553e-01 -6.22859955e-01 1.49552420e-01
7.67106175e-01 4.90828097e-01 -9.34951127e-01 5.24301529e-01
-7.20461607e-01 3.20201844e-01 1.22083694e-01 -5.66198826e-02
2.03237250e-01 1.46284103e-02 3.96330357e-01 1.02191889e+00
-2.18559295e-01 5.54560959e-01 -1.15498729e-01 1.41033280e+00
-1.93240106e-01 -5.93443155e-01 -4.67169344e-01 -7.90196955e-02
3.63906026e-01 1.12661433e+00 -4.93239194e-01 -5.96464932e-01
-4.47494179e-01 1.13192773e+00 6.48776293e-01 4.62523878e-01
-9.49829817e-01 -5.00323512e-02 9.66440856e-01 -2.26894379e-01
-1.78080633e-01 -1.89358726e-01 -5.09194732e-01 -1.54193926e+00
6.95557818e-02 -1.50948083e+00 3.92089933e-01 -1.65692377e+00
-1.43922853e+00 2.32852906e-01 -2.25131586e-01 -7.26652980e-01
-2.27191627e-01 -1.09354913e+00 -3.73567700e-01 7.96190202e-01
-1.58038449e+00 -1.15866280e+00 -7.79560864e-01 6.88971937e-01
6.53780878e-01 2.23213613e-01 8.98834109e-01 -2.00955570e-01
7.31114969e-02 2.82888293e-01 -7.23435700e-01 4.76453900e-01
7.51450956e-01 -1.37672472e+00 7.84593344e-01 9.07972097e-01
4.70055848e-01 9.41640437e-01 9.48467910e-01 -3.88956964e-01
-1.40294671e+00 -7.58527517e-01 8.02957296e-01 -1.21701479e+00
7.74923205e-01 -3.04612428e-01 -9.24470484e-01 7.76351452e-01
3.60475391e-01 -3.03148255e-02 7.12060750e-01 2.20599130e-01
-1.06805658e+00 2.09143057e-01 -1.13400781e+00 1.23509359e+00
1.30432773e+00 -1.00591397e+00 -1.42230499e+00 3.64557207e-01
9.16971087e-01 -3.85443568e-01 -5.86435080e-01 2.65310854e-01
6.22843742e-01 -1.11765647e+00 1.22950006e+00 -1.22871006e+00
8.92395437e-01 -5.72994471e-01 -5.31160116e-01 -1.04784179e+00
-5.07494271e-01 -5.38833737e-01 -7.71479532e-02 8.66680384e-01
2.68147349e-01 -2.95951009e-01 5.31641126e-01 1.30683708e+00
1.55931562e-01 -4.38627154e-01 -6.72440529e-01 -4.82429296e-01
9.69517007e-02 -8.95359457e-01 6.08925402e-01 1.11698282e+00
-2.92461086e-02 5.61919630e-01 -1.55394033e-01 2.08356142e-01
7.87331581e-01 4.10455197e-01 7.99498439e-01 -9.74452317e-01
-5.73782563e-01 -3.13073277e-01 -3.75236571e-01 -1.15582454e+00
1.86544716e-01 -8.48081589e-01 1.80711076e-02 -1.92294908e+00
5.88586628e-01 -3.60087678e-02 -1.22325830e-01 7.02896714e-01
-3.30343574e-01 4.69845891e-01 6.94404066e-01 6.62317201e-02
-9.43003953e-01 1.18427984e-02 1.68262732e+00 -4.56927508e-01
6.34805977e-01 -5.73496640e-01 -1.24297333e+00 1.16588938e+00
6.71264946e-01 -7.29892170e-03 -4.46093410e-01 -8.04741263e-01
8.08814347e-01 1.83825970e-01 1.38911927e+00 -8.68441582e-01
3.73938501e-01 -5.15063703e-01 5.06064177e-01 -3.92541856e-01
5.33116758e-01 -6.63530648e-01 -4.61884066e-02 3.32853258e-01
-5.56039095e-01 6.14595506e-03 -2.94285677e-02 4.83321309e-01
1.68372057e-02 -1.39088795e-01 7.17726469e-01 -5.38633466e-01
-1.27471173e+00 -1.89438581e-01 -1.61472023e-01 6.24750555e-01
7.07968473e-01 -2.04795748e-01 -1.06465387e+00 -6.97376966e-01
-9.34039652e-01 3.70652914e-01 5.74025810e-01 4.15375352e-01
8.65755320e-01 -1.12164915e+00 -7.54931569e-01 -3.32201630e-01
3.42858106e-01 -1.42609537e-01 4.52448517e-01 3.26509207e-01
-4.57617491e-01 3.81853074e-01 -3.84088397e-01 -3.60258311e-01
-1.11836028e+00 7.74730027e-01 4.10171986e-01 3.48514766e-02
-5.56614459e-01 9.61098373e-01 3.83173078e-01 -2.64906287e-01
-2.13318374e-02 -4.21081007e-01 -1.96908593e-01 -2.91038185e-01
7.85384059e-01 1.34907186e-01 -3.59619975e-01 -5.39478064e-01
-4.44801927e-01 3.92593384e-01 1.81129500e-01 -2.12760791e-01
9.25037265e-01 -1.13641873e-01 -3.61017436e-01 2.71339536e-01
6.18790925e-01 -3.59124810e-01 -1.29249811e+00 -1.30401373e-01
-2.34585837e-01 -6.25849962e-01 -4.13839608e-01 -1.28091264e+00
-3.84567678e-01 1.24467301e+00 1.84635699e-01 5.18720746e-02
1.03291106e+00 2.44506776e-01 5.07806659e-01 8.76501083e-01
4.87083226e-01 -8.90899420e-01 5.29247642e-01 6.36850238e-01
1.16271734e+00 -1.70712996e+00 -1.00355700e-01 -3.44494849e-01
-8.96252334e-01 9.31963801e-01 9.31424439e-01 -1.58285588e-01
8.02454650e-02 -1.27924681e-01 1.89395458e-01 -4.38817799e-01
-1.01200128e+00 -3.74836028e-01 4.30656701e-01 8.45499754e-01
1.13700099e-01 2.00967222e-01 3.10887039e-01 7.45021939e-01
-4.47132736e-01 -8.93400237e-02 5.02546132e-01 3.99541169e-01
-4.52718675e-01 -4.82554764e-01 -4.64287013e-01 2.13416845e-01
-2.50016153e-01 -3.35849553e-01 -6.38376713e-01 5.90505242e-01
3.79001945e-01 1.30935371e+00 -1.27324462e-03 -2.86667973e-01
1.72627985e-01 1.15054935e-01 8.34178090e-01 -5.77565491e-01
-3.44258398e-01 -9.12269831e-01 3.92945349e-01 -8.95697594e-01
-4.98725265e-01 -3.56051326e-01 -1.22459888e+00 -3.48847419e-01
2.53117621e-01 -3.21531713e-01 3.66753727e-01 1.27592945e+00
2.52467424e-01 5.27568161e-01 -2.28929624e-01 -3.97878557e-01
-5.25943100e-01 -5.02668142e-01 1.85117275e-01 9.97699738e-01
3.00987750e-01 -4.18387294e-01 -3.25969875e-01 4.42540824e-01]
|
[10.786489486694336, 1.937195897102356]
|
6744747c-897b-4825-a602-f26734e39be5
|
cross-domain-aspect-extraction-using
|
2210.10144
| null |
https://arxiv.org/abs/2210.10144v1
|
https://arxiv.org/pdf/2210.10144v1.pdf
|
Cross-Domain Aspect Extraction using Transformers Augmented with Knowledge Graphs
|
The extraction of aspect terms is a critical step in fine-grained sentiment analysis of text. Existing approaches for this task have yielded impressive results when the training and testing data are from the same domain. However, these methods show a drastic decrease in performance when applied to cross-domain settings where the domain of the testing data differs from that of the training data. To address this lack of extensibility and robustness, we propose a novel approach for automatically constructing domain-specific knowledge graphs that contain information relevant to the identification of aspect terms. We introduce a methodology for injecting information from these knowledge graphs into Transformer models, including two alternative mechanisms for knowledge insertion: via query enrichment and via manipulation of attention patterns. We demonstrate state-of-the-art performance on benchmark datasets for cross-domain aspect term extraction using our approach and investigate how the amount of external knowledge available to the Transformer impacts model performance.
|
['Gadi Singer', 'Moshe Wasserblat', 'Oren Pereg', 'Daniel Korat', 'Ana Paula Simoes', 'Vasudev Lal', 'Arden Ma', 'Phillip Howard']
|
2022-10-18
| null | null | null | null |
['term-extraction', 'aspect-extraction']
|
['natural-language-processing', 'natural-language-processing']
|
[ 3.62392932e-01 1.84892997e-01 -4.22249556e-01 -4.67140973e-01
-1.08677185e+00 -9.95773315e-01 8.04366410e-01 4.25299197e-01
-2.93981791e-01 5.13391554e-01 8.09227526e-02 -4.84047830e-01
-1.06953070e-01 -1.03573501e+00 -7.63159692e-01 -1.42087445e-01
3.26262325e-01 7.11809099e-01 4.20537740e-01 -4.83438075e-01
2.93924361e-01 2.19894141e-01 -1.56586814e+00 4.58038002e-01
8.10205400e-01 9.84001577e-01 -3.36093068e-01 2.70934522e-01
-5.15294731e-01 4.26931679e-01 -8.28892767e-01 -8.04871976e-01
1.46227688e-01 -9.28714722e-02 -1.05449069e+00 9.29153264e-02
3.90416473e-01 1.78291366e-01 2.34181941e-01 1.00605726e+00
3.59316945e-01 -1.33898094e-01 6.40805185e-01 -1.26235604e+00
-6.68396950e-01 7.79612839e-01 -4.41117615e-01 2.67319381e-01
3.50443006e-01 -1.83317021e-01 1.45181549e+00 -7.55805969e-01
7.20946074e-01 9.24388826e-01 7.56216586e-01 4.47057694e-01
-1.20777452e+00 -5.20281792e-01 3.67589980e-01 5.56915477e-02
-1.18028808e+00 -4.90687311e-01 7.16276526e-01 -2.61318743e-01
1.63797104e+00 1.05205454e-01 5.30924737e-01 8.39990079e-01
-2.02551737e-01 6.32753313e-01 9.57276583e-01 -6.59747541e-01
2.06224278e-01 6.62808239e-01 3.99510741e-01 5.42387545e-01
5.76352358e-01 -1.26417473e-01 -7.17081547e-01 -5.27257383e-01
2.98741788e-01 -5.90244949e-01 -1.02321096e-01 -4.73849475e-01
-7.41787791e-01 9.58610535e-01 -4.41311449e-02 4.14646983e-01
-2.72927672e-01 -1.87030837e-01 5.19193292e-01 4.40823436e-01
5.98717570e-01 1.05471587e+00 -1.18354356e+00 -1.01133794e-01
-8.30599308e-01 1.96321577e-01 1.15715599e+00 1.24594712e+00
7.51119316e-01 -2.82319427e-01 -1.36618108e-01 6.85349941e-01
5.43984845e-02 3.89781445e-01 6.09412193e-01 -4.15891558e-01
7.12774038e-01 1.20451045e+00 2.22557336e-02 -6.87295973e-01
-2.04284281e-01 -6.12302244e-01 4.74407449e-02 -2.61347443e-01
2.84968108e-01 -1.46136686e-01 -1.00190604e+00 1.72606409e+00
6.32790387e-01 -3.90169919e-01 1.78584084e-01 2.54727304e-01
8.34881783e-01 1.28952891e-01 1.10162638e-01 1.67950481e-01
1.68318295e+00 -9.70152140e-01 -4.68698710e-01 -6.28917158e-01
8.61736715e-01 -6.83500171e-01 1.31780517e+00 1.87454939e-01
-8.40142310e-01 -1.24736279e-01 -1.10923815e+00 -2.67541278e-02
-9.37526643e-01 -3.58732305e-02 6.16410673e-01 8.32404435e-01
-6.96138501e-01 4.00183409e-01 -4.58832473e-01 -4.04436409e-01
3.84330660e-01 4.61539894e-01 -3.83380204e-01 -9.96293351e-02
-1.26515985e+00 1.00769794e+00 2.66910523e-01 -5.08473694e-01
-3.82910252e-01 -1.05185807e+00 -9.34543967e-01 2.48273849e-01
5.43280780e-01 -8.62414539e-01 1.32780826e+00 -1.15588653e+00
-8.72732103e-01 1.03390396e+00 -2.93076187e-01 -2.81257719e-01
-2.96604838e-02 -2.87512213e-01 -5.40608823e-01 -6.56930134e-02
4.84186202e-01 1.45921826e-01 9.50537682e-01 -1.11077559e+00
-7.18968511e-01 -5.73852599e-01 5.16698897e-01 1.64913625e-01
-5.90758204e-01 1.72936901e-01 -7.56754160e-01 -6.46537960e-01
-3.69632393e-01 -7.91995406e-01 -5.94219752e-02 -5.18724322e-01
-4.48626846e-01 -2.07122341e-01 8.10651124e-01 -3.70862424e-01
1.48883617e+00 -1.93450058e+00 -1.47045419e-01 4.28779721e-01
1.27300784e-01 3.18787158e-01 -1.60670429e-01 3.54590684e-01
-7.66279548e-02 4.75777507e-01 -1.11784086e-01 -1.86616629e-01
1.23070717e-01 6.50790110e-02 -3.93674970e-01 -2.56504685e-01
5.46867132e-01 1.00070596e+00 -7.62607217e-01 -4.36322361e-01
-3.79483193e-01 3.47865582e-01 -6.58818483e-01 1.22625358e-01
-5.19806266e-01 -8.82651210e-02 -6.84390903e-01 7.23956466e-01
3.80918682e-01 -5.61172783e-01 3.02739799e-01 -4.99288350e-01
4.07995194e-01 8.80965173e-01 -1.03258646e+00 1.22199404e+00
-6.12097144e-01 3.82853866e-01 -2.02634558e-01 -7.07569301e-01
6.07527614e-01 2.27916956e-01 3.04289848e-01 -7.38910913e-01
4.21646237e-02 6.35687709e-02 -1.91300198e-01 -3.34254533e-01
6.15624428e-01 -5.19768715e-01 -3.19727242e-01 7.42444754e-01
3.39484096e-01 -3.64660561e-01 5.38684309e-01 1.13249049e-01
1.31068039e+00 -7.23940954e-02 5.26032329e-01 -2.03642949e-01
4.47831422e-01 5.25562048e-01 3.71711224e-01 6.26929283e-01
1.51956871e-01 1.44131467e-01 5.80854356e-01 -2.39642128e-01
-8.98867548e-01 -6.61881089e-01 -9.12048146e-02 1.27626646e+00
-1.77591637e-01 -9.26607609e-01 -7.40967631e-01 -1.32799816e+00
2.13129699e-01 8.05992067e-01 -8.13487053e-01 -4.14433599e-01
-4.13116336e-01 -8.88935089e-01 6.81173205e-01 7.18145251e-01
2.56228566e-01 -8.86967301e-01 -4.87769157e-01 1.78486213e-01
-2.28767306e-01 -1.52194250e+00 -2.38370851e-01 5.22296011e-01
-8.21866095e-01 -1.30129635e+00 -4.74585742e-02 -6.21920705e-01
7.97683120e-01 6.12040944e-02 2.01884365e+00 1.17452890e-02
-1.31429583e-01 5.74001074e-01 -4.71471310e-01 -6.00531399e-01
-3.42747241e-01 4.53890681e-01 -3.51657659e-01 -2.98942894e-01
1.10240221e+00 -3.33918899e-01 -3.21889043e-01 4.09303695e-01
-1.05074978e+00 -3.56579632e-01 5.78711450e-01 7.57080734e-01
7.36662209e-01 3.72476101e-01 5.41979671e-01 -1.50743437e+00
1.01467693e+00 -4.12260562e-01 -6.69252276e-01 3.85067910e-01
-9.93496716e-01 4.02338505e-01 5.71315229e-01 -3.86122972e-01
-9.74143624e-01 -8.55637938e-02 -5.75642241e-03 -2.11039651e-03
-1.06963366e-01 7.84089983e-01 -2.62880057e-01 -1.98580801e-01
6.96693957e-01 1.27247768e-02 -4.71605092e-01 -4.68133807e-01
4.02420461e-01 4.09837455e-01 -5.62144108e-02 -7.87957907e-01
7.63813078e-01 3.99373233e-01 -2.28319779e-01 -5.75222969e-01
-1.19183338e+00 -6.36757851e-01 -5.66555440e-01 4.56308573e-01
5.93160093e-01 -8.50785255e-01 -4.72340956e-02 1.68873832e-01
-9.05511975e-01 -3.62717099e-02 -6.74364686e-01 -4.40730713e-03
-3.54510933e-01 8.56568888e-02 -2.48120651e-01 -3.10223609e-01
-5.89031279e-01 -1.07103205e+00 1.35700214e+00 1.63487960e-02
-4.82767999e-01 -1.21492469e+00 2.10905880e-01 4.73998517e-01
3.71600181e-01 -1.63867921e-01 1.50027633e+00 -1.15832663e+00
-4.64420527e-01 -3.92373532e-01 -1.65996850e-01 3.09168786e-01
4.00968939e-01 -2.62025952e-01 -1.06760967e+00 -3.20694707e-02
-4.92432453e-02 -2.25816563e-01 5.09311318e-01 -3.47968265e-02
5.97126722e-01 -3.74273330e-01 -5.05850196e-01 3.71539801e-01
1.49591434e+00 -1.82667673e-02 2.98064619e-01 7.86883533e-01
6.75450027e-01 6.33965552e-01 7.95526147e-01 2.29932651e-01
4.91874009e-01 7.37032413e-01 2.67925430e-02 -1.13529712e-01
-8.90960023e-02 -2.01409742e-01 2.34562516e-01 4.77900088e-01
2.85084993e-01 -2.23241702e-01 -9.59485292e-01 1.03313649e+00
-1.49436545e+00 -7.12069213e-01 2.67555505e-01 2.12133098e+00
1.21111333e+00 4.32899147e-01 7.51172453e-02 2.00322971e-01
4.00560200e-01 -3.16203050e-02 -5.76093853e-01 -4.97305095e-01
-1.33572996e-01 6.16164207e-01 5.50489664e-01 2.62755126e-01
-1.21334004e+00 1.10753679e+00 6.82478380e+00 6.91994011e-01
-9.99992132e-01 -4.25632019e-03 1.37780458e-01 -6.16484433e-02
-6.62499011e-01 1.15845911e-01 -1.18930459e+00 3.78589556e-02
9.36290741e-01 -3.42645943e-01 1.28724992e-01 8.65041256e-01
-3.21168959e-01 3.67748290e-02 -1.25548029e+00 3.98184121e-01
1.40031502e-01 -1.17232108e+00 3.86642575e-01 -8.22441950e-02
7.16529310e-01 9.39851254e-02 5.77623770e-03 4.11143988e-01
4.21283215e-01 -8.38022530e-01 2.75045186e-01 -5.54815866e-02
7.24143445e-01 -7.61899471e-01 6.82795525e-01 5.52265421e-02
-1.21985507e+00 3.16763557e-02 4.00415994e-02 2.57010102e-01
-6.11377917e-02 7.15698183e-01 -1.18949342e+00 5.10934770e-01
7.60505140e-01 4.60601836e-01 -9.62541223e-01 6.69414580e-01
-3.51620317e-01 5.59925556e-01 -2.68109381e-01 2.31763325e-03
1.25163496e-01 2.47337833e-01 5.33155859e-01 1.30877960e+00
4.11336496e-02 -4.27279919e-01 -1.91732168e-01 6.48834050e-01
-1.72545344e-01 2.44361833e-01 -8.99917603e-01 -3.79358321e-01
3.32935631e-01 1.17176044e+00 -5.80317438e-01 -5.35160244e-01
-7.53193498e-01 5.59310436e-01 4.88420397e-01 3.62449050e-01
-6.14226341e-01 -6.09191716e-01 8.75412822e-01 2.41073146e-01
9.41930056e-01 3.16714458e-02 -4.18558747e-01 -1.35174012e+00
4.31367666e-01 -1.37061191e+00 7.24339902e-01 -4.63742733e-01
-1.39606106e+00 7.48759449e-01 1.42677620e-01 -9.79596615e-01
-3.90643388e-01 -6.47486329e-01 -3.57185036e-01 8.63386691e-01
-1.75570917e+00 -1.28247833e+00 -6.10209145e-02 6.20203316e-01
3.28915566e-01 -7.41781816e-02 1.07783365e+00 4.46787179e-01
-1.19491354e-01 9.61496115e-01 -1.56043708e-01 2.07697481e-01
7.47819006e-01 -1.43734145e+00 8.11630785e-01 8.65186036e-01
2.76664704e-01 1.04851770e+00 6.94796979e-01 -7.21745968e-01
-1.40843356e+00 -1.19568241e+00 1.15011501e+00 -9.56973612e-01
8.92517745e-01 -4.66051728e-01 -9.36064243e-01 9.32946622e-01
1.51887015e-01 -2.93025643e-01 1.14490402e+00 6.96223378e-01
-8.26519608e-01 7.15900538e-03 -1.19501340e+00 4.44436818e-01
9.03655648e-01 -8.36896956e-01 -9.56426322e-01 1.30881920e-01
7.97621906e-01 -2.20692396e-01 -1.02477479e+00 5.55704176e-01
4.48015600e-01 -5.70574701e-01 8.54570150e-01 -1.06770158e+00
3.64493698e-01 -3.45313340e-01 -2.14578763e-01 -1.36624062e+00
-4.43653651e-02 -3.21785092e-01 -2.69368291e-01 1.43487751e+00
1.06973195e+00 -6.88323677e-01 8.41099322e-01 7.08133340e-01
2.49464571e-01 -8.46739173e-01 -6.26116872e-01 -6.25510931e-01
8.84316191e-02 -4.41891015e-01 9.86389518e-01 9.58055615e-01
4.68645208e-02 8.76816332e-01 1.87828571e-01 1.33924291e-01
3.71034533e-01 5.43221176e-01 7.18895376e-01 -1.12654424e+00
-4.45857227e-01 -3.87175411e-01 -3.45012516e-01 -6.99914038e-01
1.46173894e-01 -6.93394065e-01 -1.69381082e-01 -1.51052308e+00
1.63732931e-01 -5.11217415e-01 -2.69048035e-01 6.31728888e-01
-3.14488053e-01 1.40670270e-01 -1.27780929e-01 -1.80200666e-01
-5.95578849e-01 2.86882460e-01 1.01485610e+00 -3.03416014e-01
-8.95392224e-02 1.10265300e-01 -1.27306914e+00 6.46779299e-01
6.30941689e-01 -7.05936372e-01 -6.77776992e-01 -4.73778099e-01
7.10128248e-01 -5.35482049e-01 -3.75417694e-02 -5.08524656e-01
1.34046346e-01 9.39613730e-02 1.23158559e-01 -3.19297314e-01
2.55256265e-01 -9.59394813e-01 -2.87073404e-01 7.72178173e-03
-2.35408068e-01 4.45733905e-01 5.52727878e-01 5.30424356e-01
-5.42027950e-01 -2.82274365e-01 5.03252864e-01 -2.60354102e-01
-6.73174381e-01 8.84350538e-02 -2.02632219e-01 6.96712732e-01
7.98519075e-01 2.35800948e-02 -6.33189499e-01 -1.32985055e-01
-5.10521889e-01 4.22287993e-02 6.22380257e-01 5.80884874e-01
2.87887812e-01 -1.10951591e+00 -2.39090502e-01 3.37350905e-01
6.27455533e-01 -1.77187324e-01 -2.82266498e-01 5.98639250e-01
5.11801094e-02 5.65283835e-01 4.71930131e-02 -2.40428120e-01
-1.42119229e+00 7.47466862e-01 3.38230968e-01 -9.17389929e-01
-1.88643187e-01 7.07036257e-01 2.47253418e-01 -7.33092725e-01
-2.92618107e-02 -3.45262796e-01 -3.09549034e-01 2.95324177e-01
3.46271813e-01 -9.13300440e-02 7.77840197e-01 -5.08964539e-01
-6.29620850e-01 6.84625924e-01 -3.55821311e-01 -1.52925089e-01
1.29814744e+00 1.59316510e-02 -8.30200166e-02 2.28143990e-01
1.10098243e+00 3.98120850e-01 -3.99692804e-01 -4.38236505e-01
3.76857847e-01 -3.73623729e-01 -3.34487706e-02 -1.15364754e+00
-9.83881354e-01 5.88765264e-01 1.16461925e-01 4.33708787e-01
1.29916632e+00 2.37280533e-01 6.87187612e-01 4.66944963e-01
2.95897722e-01 -1.13028967e+00 -1.19930945e-01 7.58739054e-01
5.60751557e-01 -1.29485333e+00 1.03134051e-01 -5.52383184e-01
-6.35871530e-01 7.83354282e-01 6.49540961e-01 1.47026360e-01
7.59760201e-01 4.66792732e-01 2.77592719e-01 -6.59907281e-01
-9.83396053e-01 -4.76160854e-01 5.86793065e-01 6.83694959e-01
5.73229194e-01 -1.87933862e-01 -7.06561506e-02 8.91934395e-01
-2.89255977e-01 -1.85934320e-01 1.71403214e-01 1.37182510e+00
-1.39788002e-01 -1.46413016e+00 -5.78077231e-03 6.87469125e-01
-8.59044790e-01 -6.13126099e-01 -8.51752460e-01 8.37186933e-01
-6.06127121e-02 9.30462956e-01 -3.71387988e-01 -3.34724545e-01
7.17215300e-01 3.19842786e-01 4.65873808e-01 -1.02566850e+00
-1.08820856e+00 -1.86335370e-01 7.66179800e-01 -5.14766157e-01
-4.13342923e-01 -4.95437711e-01 -9.67323661e-01 1.09698378e-01
-5.70077538e-01 5.40720105e-01 6.23537958e-01 1.20356905e+00
8.23473930e-01 5.64289093e-01 2.96838671e-01 -5.72263189e-02
-3.68481666e-01 -6.96780741e-01 -2.88555950e-01 4.99022156e-01
2.83103228e-01 -7.20133483e-01 -2.54900873e-01 1.65931627e-01]
|
[11.327338218688965, 6.852653980255127]
|
221f3f41-da56-4a23-8551-dea0d9a004df
|
4dac-learning-attribute-compression-for
|
2204.11723
| null |
https://arxiv.org/abs/2204.11723v1
|
https://arxiv.org/pdf/2204.11723v1.pdf
|
4DAC: Learning Attribute Compression for Dynamic Point Clouds
|
With the development of the 3D data acquisition facilities, the increasing scale of acquired 3D point clouds poses a challenge to the existing data compression techniques. Although promising performance has been achieved in static point cloud compression, it remains under-explored and challenging to leverage temporal correlations within a point cloud sequence for effective dynamic point cloud compression. In this paper, we study the attribute (e.g., color) compression of dynamic point clouds and present a learning-based framework, termed 4DAC. To reduce temporal redundancy within data, we first build the 3D motion estimation and motion compensation modules with deep neural networks. Then, the attribute residuals produced by the motion compensation component are encoded by the region adaptive hierarchical transform into residual coefficients. In addition, we also propose a deep conditional entropy model to estimate the probability distribution of the transformed coefficients, by incorporating temporal context from consecutive point clouds and the motion estimation/compensation modules. Finally, the data stream is losslessly entropy coded with the predicted distribution. Extensive experiments on several public datasets demonstrate the superior compression performance of the proposed approach.
|
['Yulan Guo', 'Yiling Xu', 'Qingyong Hu', 'Guangchi Fang']
|
2022-04-25
| null | null | null | null |
['motion-compensation']
|
['computer-vision']
|
[ 2.38690302e-01 -4.60472435e-01 -1.63364097e-01 -2.77983010e-01
-6.08799875e-01 -1.34090930e-01 3.48021686e-01 1.18453734e-01
-1.52932301e-01 2.10429952e-01 1.62347242e-01 1.46309147e-02
-1.71360195e-01 -8.59266698e-01 -8.81924272e-01 -7.63494551e-01
-3.90682518e-01 3.89241457e-01 2.23226458e-01 2.11805001e-01
4.50570583e-01 8.68426442e-01 -1.71092892e+00 2.04420239e-01
8.75941098e-01 1.39998925e+00 5.37339687e-01 7.38858998e-01
-2.49431476e-01 6.51032388e-01 -3.20780635e-01 -7.15131313e-02
5.23984849e-01 2.54754405e-02 -5.26302516e-01 2.74623007e-01
2.42563501e-01 -8.99886847e-01 -6.35206223e-01 9.27347362e-01
3.77438426e-01 1.70142636e-01 3.63107324e-01 -1.16112518e+00
-3.69265169e-01 -2.20929272e-02 -5.84565520e-01 7.85588026e-02
3.67556885e-02 2.58378625e-01 7.05809414e-01 -9.83237207e-01
6.50665939e-01 1.24311841e+00 5.24025023e-01 2.18682691e-01
-8.82957816e-01 -7.01530337e-01 -8.64119455e-02 4.89639640e-01
-1.45623541e+00 -1.58751667e-01 9.48237658e-01 -4.04863656e-01
1.18467307e+00 1.84744924e-01 1.02638030e+00 3.53669047e-01
1.86372012e-01 6.89902306e-01 5.91547191e-01 3.56692560e-02
2.38582015e-01 -4.06447321e-01 -5.32989621e-01 3.70636493e-01
2.41554454e-02 3.24539930e-01 -5.33899069e-01 -8.78791809e-02
6.30168974e-01 5.28048515e-01 -3.03908974e-01 -4.51472759e-01
-9.96100366e-01 5.14642894e-01 6.38017654e-01 -9.67847183e-02
-6.04022205e-01 5.45546234e-01 3.75462085e-01 1.81050733e-01
7.52844334e-01 -2.22015783e-01 -3.89572769e-01 -2.38938078e-01
-1.36541998e+00 3.28632385e-01 3.69124740e-01 1.30611408e+00
7.76909709e-01 9.33020785e-02 -2.62734387e-02 7.43457496e-01
4.82043713e-01 6.94824517e-01 2.06290781e-01 -1.31463683e+00
7.18740463e-01 5.21721840e-01 -6.20868169e-02 -1.23233879e+00
7.15206638e-02 -2.03699380e-01 -1.11547828e+00 2.66986609e-01
-5.56591868e-01 4.63737816e-01 -7.85590529e-01 1.26387775e+00
5.27163148e-01 8.08541715e-01 3.79129895e-04 8.16249669e-01
2.60546297e-01 1.02454019e+00 -4.31181155e-02 -3.69834512e-01
7.73783267e-01 -6.24848247e-01 -6.47636592e-01 2.01706514e-01
4.02070284e-01 -7.00149238e-01 5.37245512e-01 2.49577463e-01
-1.42073977e+00 -5.52000105e-01 -1.18031919e+00 -3.54628980e-01
3.59537378e-02 -5.10672033e-01 3.14231604e-01 2.69907359e-02
-8.81248832e-01 9.64449108e-01 -1.23062932e+00 1.48342714e-01
6.69952035e-01 3.83451253e-01 -8.03847760e-02 -4.37623292e-01
-8.88569474e-01 4.97894228e-01 4.71902758e-01 -3.07542849e-02
-7.24767864e-01 -8.69809568e-01 -7.21918762e-01 2.59663522e-01
-1.56230107e-01 -7.45513380e-01 1.12948346e+00 -3.55955869e-01
-1.05393529e+00 5.17240703e-01 -3.16051632e-01 -6.55430675e-01
4.42992628e-01 -3.59796047e-01 -1.23504780e-01 4.68332767e-01
-6.73621297e-02 9.99466300e-01 8.82906318e-01 -1.36121023e+00
-9.99956429e-01 -3.79057974e-01 -5.45021117e-01 3.77147764e-01
-2.05458537e-01 -3.44716996e-01 -9.07129109e-01 -6.86078966e-01
4.73345011e-01 -9.59346652e-01 -2.29796797e-01 4.78582293e-01
1.05263144e-01 2.00745594e-02 1.30365300e+00 -8.71085525e-01
1.24522352e+00 -2.55956364e+00 1.20101541e-01 1.42883077e-01
1.84300825e-01 4.29161591e-03 1.31549509e-02 1.16277762e-01
1.03749009e-02 1.24184132e-01 -7.59117723e-01 -5.83275676e-01
-4.86844592e-02 4.74333227e-01 -6.66399062e-01 4.36552346e-01
3.23409110e-01 6.10432267e-01 -7.60331690e-01 -6.69692934e-01
5.88642955e-01 7.74804175e-01 -8.59620690e-01 2.40672812e-01
-2.84149170e-01 2.54258811e-01 -3.63470972e-01 8.65708411e-01
1.34875989e+00 -2.50563174e-01 -2.96220243e-01 -1.12906955e-01
-1.46835059e-01 1.68169484e-01 -8.05338800e-01 1.92239869e+00
-2.30949372e-01 6.65010750e-01 -7.36612976e-02 -6.33268833e-01
8.68524492e-01 2.14598134e-01 1.17397285e+00 -6.17483318e-01
-1.82874829e-01 3.15606356e-01 -3.77773434e-01 -3.32800865e-01
1.08049905e+00 -8.68376642e-02 9.43756774e-02 5.25341108e-02
-3.00044358e-01 -6.75965488e-01 -3.11864674e-01 1.32596374e-01
1.02306247e+00 1.97983071e-01 -8.63786191e-02 5.89100063e-01
3.67813468e-01 1.62591860e-01 6.36749566e-01 1.87449619e-01
-2.41882145e-01 8.57828319e-01 -7.48594105e-02 -3.69070441e-01
-1.60563207e+00 -9.12919283e-01 -1.74186543e-01 3.54053199e-01
5.27276933e-01 -3.62163633e-01 -3.21386874e-01 -1.21612854e-01
2.33963504e-01 5.56997597e-01 -4.97261472e-02 -4.11831260e-01
-8.07893395e-01 -3.02912921e-01 1.93825811e-01 4.36661512e-01
7.56400764e-01 -7.56632984e-01 -7.07485497e-01 2.91189492e-01
-3.74593019e-01 -1.20074534e+00 -4.39876080e-01 -8.59247521e-02
-1.57193851e+00 -5.01630127e-01 -4.01178807e-01 -4.70677674e-01
3.15641761e-01 6.77353740e-01 1.07984364e+00 3.72418791e-01
-2.72973883e-03 3.45414370e-01 -5.19603908e-01 -3.13802421e-01
-2.44922817e-01 -2.32612506e-01 -2.43920431e-01 -3.37938100e-01
3.82688731e-01 -9.28541481e-01 -9.39219534e-01 -3.27293128e-02
-1.26137853e+00 1.05804145e-01 7.58988321e-01 4.33582544e-01
1.20639801e+00 2.95744628e-01 -1.89851001e-01 -2.56301433e-01
1.69550806e-01 -7.24852264e-01 -6.02121651e-01 -1.72251716e-01
-7.23553240e-01 3.94304991e-02 2.43661910e-01 -5.57636768e-02
-7.57517338e-01 2.90341228e-01 -2.82350272e-01 -1.25881708e+00
1.05253987e-01 4.14697319e-01 -8.58457312e-02 -5.27668186e-02
-1.36173651e-01 5.56093454e-01 -1.42726079e-01 -3.94116849e-01
2.45628878e-01 6.10764444e-01 8.64017189e-01 -3.12830061e-01
1.04575026e+00 7.96844602e-01 2.11457253e-01 -6.33748293e-01
-2.68402159e-01 -5.16884685e-01 -6.30045652e-01 -3.51576924e-01
9.67155814e-01 -1.17415512e+00 -5.67038238e-01 3.48593235e-01
-1.44853234e+00 8.89854878e-02 -4.31677848e-01 5.25258660e-01
-8.02418888e-01 7.54675865e-01 -6.34313881e-01 -7.22555578e-01
-6.39245093e-01 -1.13426900e+00 1.37864971e+00 -1.59579083e-01
2.17788592e-01 -4.74379450e-01 4.52577174e-02 1.39365196e-01
2.47966111e-01 3.98703665e-01 9.53237951e-01 -1.82056054e-02
-1.56182230e+00 -3.08443278e-01 -2.13708043e-01 4.10691559e-01
-7.41090029e-02 1.33264467e-01 -6.65939212e-01 -3.03943276e-01
3.55532169e-01 4.49647233e-02 8.23140800e-01 4.28289592e-01
1.71974063e+00 -3.51167947e-01 -2.92148203e-01 1.28216636e+00
1.54485679e+00 2.82608300e-01 7.85166085e-01 2.94191390e-01
6.85231447e-01 2.32064962e-01 8.98713052e-01 1.02328312e+00
4.71930027e-01 5.00365496e-01 9.67208028e-01 4.44844782e-01
-9.22435299e-02 -3.84729058e-01 1.84643731e-01 1.54572070e+00
-1.29483476e-01 -2.55544305e-01 -7.77174711e-01 4.59295422e-01
-1.59805703e+00 -1.08214331e+00 1.12635177e-02 2.13640523e+00
5.69154322e-01 -2.03456134e-02 -6.00931466e-01 2.34776512e-01
6.85515821e-01 4.29502904e-01 -8.48239481e-01 -1.41842142e-01
-1.60882130e-01 1.00176983e-01 6.35057032e-01 3.03030699e-01
-1.03255248e+00 6.12883866e-01 5.68996382e+00 7.67292380e-01
-1.16049969e+00 -4.19669375e-02 3.94231170e-01 -4.41377491e-01
-4.24965471e-01 1.54130116e-01 -5.53004026e-01 8.45813751e-01
1.03515959e+00 -2.58592218e-01 3.86364698e-01 1.01780570e+00
2.17862546e-01 1.32348195e-01 -8.91428292e-01 1.26276946e+00
-1.16633005e-01 -1.46091795e+00 2.73105294e-01 4.01490986e-01
7.25182891e-01 3.29654425e-01 2.09226325e-01 1.37813836e-01
6.73383549e-02 -6.52657509e-01 8.53119433e-01 6.44248128e-01
9.73644614e-01 -9.43093836e-01 4.48893934e-01 3.28158796e-01
-1.43489528e+00 -1.77920967e-01 -7.88639009e-01 2.62492359e-01
3.79890949e-01 6.88943386e-01 -8.65112424e-01 7.80134201e-01
9.39788699e-01 1.17772436e+00 -2.63784558e-01 1.25953388e+00
2.50878900e-01 3.42404813e-01 -4.98766303e-01 4.91985947e-01
1.46076858e-01 -2.17860550e-01 7.14103639e-01 9.86716151e-01
9.75911081e-01 3.50233406e-01 -1.49014726e-01 6.75519407e-01
-2.14997917e-01 -1.96338147e-01 -6.89723611e-01 1.23697199e-01
8.27484548e-01 7.86016107e-01 -3.66959333e-01 -2.37104177e-01
-4.62391585e-01 1.11654520e+00 1.11735925e-01 1.65412337e-01
-7.10166335e-01 6.26128316e-02 8.77777994e-01 -1.45496055e-02
6.29234076e-01 -6.13684952e-01 -4.84219253e-01 -1.03007758e+00
2.10876957e-01 -4.69101369e-01 2.64154404e-01 -9.88836467e-01
-1.05444479e+00 2.61156321e-01 1.14130937e-01 -2.04613614e+00
-2.42128998e-01 -1.64380372e-01 -4.07326341e-01 7.32907355e-01
-1.78102148e+00 -7.33256519e-01 -6.13274634e-01 5.86203933e-01
7.02877879e-01 7.53369555e-02 3.53319377e-01 6.23890340e-01
-7.81605914e-02 1.24158420e-01 3.30993980e-01 -3.22471380e-01
4.03382748e-01 -7.65471339e-01 6.30299389e-01 9.05825853e-01
-3.20132911e-01 1.90001756e-01 5.55496335e-01 -9.27931845e-01
-1.77588534e+00 -1.39638448e+00 5.67290723e-01 3.13432254e-02
3.02342653e-01 -5.52063994e-02 -1.36026967e+00 3.65804225e-01
-6.19509593e-02 1.88427240e-01 4.04151708e-01 -7.50530422e-01
-7.07217827e-02 -1.87964931e-01 -1.17083526e+00 2.37998769e-01
1.08906245e+00 -5.21434247e-01 -3.25412095e-01 1.80492878e-01
1.38495255e+00 -5.40835202e-01 -9.93967354e-01 8.02413881e-01
2.79246867e-01 -8.67849350e-01 1.18114305e+00 -1.39553607e-01
9.69038188e-01 -4.62164819e-01 -6.60939157e-01 -8.29368651e-01
-3.91975731e-01 -1.50341600e-01 -6.64321244e-01 1.08656144e+00
-1.88925251e-01 8.80132169e-02 9.53747332e-01 7.57584631e-01
-5.83005190e-01 -7.26836443e-01 -1.25465107e+00 -5.85539639e-01
-7.27338716e-02 -8.03609669e-01 9.24187183e-01 8.33304703e-01
-5.95866561e-01 -3.55526000e-01 -5.13792634e-01 5.09717822e-01
6.43515706e-01 3.83070409e-01 7.19051361e-01 -1.03161788e+00
-1.46305844e-01 -3.06226462e-01 -6.68032169e-01 -1.47352397e+00
1.88065767e-02 -9.33621228e-01 2.35013023e-01 -1.38279343e+00
8.35564509e-02 -4.79050249e-01 -1.88583896e-01 -8.19667503e-02
-6.87392876e-02 -1.31784871e-01 4.42229301e-01 8.23804140e-01
-4.72169220e-01 1.14261854e+00 1.12651396e+00 -2.99694568e-01
-8.44933614e-02 -2.43351877e-01 -3.11102741e-03 4.11158860e-01
5.82323849e-01 -6.20532632e-01 -3.45092267e-01 -8.50298464e-01
7.57715851e-02 4.35093462e-01 3.60700756e-01 -1.34032822e+00
4.97612745e-01 -2.12924376e-01 5.37535906e-01 -1.57065344e+00
6.43380940e-01 -1.38437748e+00 4.11305040e-01 5.28906047e-01
-2.25764383e-02 3.65008175e-01 1.14629298e-01 9.14797008e-01
-5.29383838e-01 1.28729036e-02 7.96985388e-01 -1.93732381e-02
-7.67642379e-01 1.03799546e+00 1.09009437e-01 -4.22898382e-01
9.37797308e-01 -4.64139670e-01 1.12988628e-01 -2.78439641e-01
-2.68327236e-01 3.09366614e-01 8.40853930e-01 2.90132344e-01
1.31356919e+00 -1.59180868e+00 -6.13108277e-01 2.32463539e-01
1.32630868e-02 7.56320357e-01 6.96339071e-01 4.87132937e-01
-9.02249098e-01 2.39371479e-01 -9.22403112e-02 -1.01365304e+00
-1.12361753e+00 7.30109453e-01 -5.37336990e-02 -1.32177562e-01
-8.90712917e-01 5.86735904e-01 -1.50698617e-01 -7.65529126e-02
1.58264413e-01 -4.49001729e-01 2.16945872e-01 -3.55465442e-01
3.56318146e-01 3.34708720e-01 5.04948646e-02 -5.66608787e-01
-9.96310040e-02 5.64161599e-01 5.27958460e-02 -1.01766482e-01
1.66953278e+00 -3.55920106e-01 -2.56150544e-01 2.46214420e-01
1.49913847e+00 -4.02646542e-01 -1.64092088e+00 -2.48956338e-01
-1.07833229e-01 -1.06347334e+00 1.21240482e-01 -2.29036529e-02
-1.27788126e+00 9.94707942e-01 8.41516078e-01 -3.40623707e-02
1.56982219e+00 -2.01964855e-01 1.29287279e+00 1.77898884e-01
2.73136169e-01 -8.92884016e-01 -1.13262601e-01 5.56225538e-01
7.35892415e-01 -1.06768906e+00 3.20593655e-01 -4.19160694e-01
-3.30087483e-01 9.35498059e-01 3.62199366e-01 -1.65126234e-01
7.05160022e-01 2.08125755e-01 -5.18572688e-01 -9.88075882e-02
-1.12005532e+00 2.54692793e-01 -3.38889542e-03 4.91173118e-01
1.38957659e-02 -1.65437073e-01 2.97403317e-02 -3.00639272e-02
-3.14012736e-01 2.40012527e-01 2.78088242e-01 1.15278101e+00
-5.66964447e-01 -8.55577409e-01 -3.98158848e-01 4.36703652e-01
-1.10269606e-01 -1.97116584e-01 6.89145774e-02 4.35360134e-01
1.18266165e-01 5.02609015e-01 4.73326594e-01 -7.28039563e-01
2.28156343e-01 -8.11418071e-02 1.04081713e-01 -2.16852024e-01
-6.22059070e-02 8.88196528e-02 -7.19624400e-01 -5.89225233e-01
-5.48573315e-01 -8.34167063e-01 -1.39017642e+00 -6.79827690e-01
5.14268056e-02 -8.73630047e-02 9.22617495e-01 4.64691550e-01
7.36447394e-01 4.11486447e-01 1.02793741e+00 -1.21464515e+00
-3.44653577e-01 -6.31916225e-01 -4.51092273e-01 5.38361490e-01
5.97565413e-01 -3.68775129e-01 -4.31038946e-01 2.32103422e-01]
|
[8.48956298828125, -3.002126455307007]
|
742434b4-b602-4a14-b04f-a506261b8224
|
investigating-non-local-features-for-neural
|
2109.12814
| null |
https://arxiv.org/abs/2109.12814v2
|
https://arxiv.org/pdf/2109.12814v2.pdf
|
Investigating Non-local Features for Neural Constituency Parsing
|
Thanks to the strong representation power of neural encoders, neural chart-based parsers have achieved highly competitive performance by using local features. Recently, it has been shown that non-local features in CRF structures lead to improvements. In this paper, we investigate injecting non-local features into the training process of a local span-based parser, by predicting constituent n-gram non-local patterns and ensuring consistency between non-local patterns and local constituents. Results show that our simple method gives better results than the self-attentive parser on both PTB and CTB. Besides, our method achieves state-of-the-art BERT-based performance on PTB (95.92 F1) and strong performance on CTB (92.31 F1). Our parser also achieves better or competitive performance in multilingual and zero-shot cross-domain settings compared with the baseline.
|
['Yue Zhang', 'Sen yang', 'Leyang Cui']
|
2021-09-27
| null |
https://aclanthology.org/2022.acl-long.146
|
https://aclanthology.org/2022.acl-long.146.pdf
|
acl-2022-5
|
['constituency-parsing']
|
['natural-language-processing']
|
[-1.91670761e-01 2.96841472e-01 -2.70971119e-01 -7.88666368e-01
-1.49911368e+00 -5.69589496e-01 5.81989706e-01 2.68650353e-01
-5.71851790e-01 8.58562887e-01 5.67999005e-01 -6.06181920e-01
-8.52701720e-04 -7.52429187e-01 -8.52660298e-01 -4.08232868e-01
-2.30409056e-01 5.70981741e-01 2.32603133e-01 -3.80506396e-01
-1.53881237e-01 4.14651364e-01 -9.77588236e-01 4.06963319e-01
7.34900773e-01 5.76312065e-01 4.60137248e-01 7.03397095e-01
-1.18789956e-01 6.77215934e-01 -5.97019136e-01 -5.97094059e-01
-7.12125972e-02 -3.23310941e-01 -1.12162280e+00 -5.78230083e-01
7.23623812e-01 -4.20572534e-02 -1.38278872e-01 7.71451592e-01
5.29664934e-01 9.41427052e-02 3.81153643e-01 -3.84416699e-01
-7.67335057e-01 1.16421270e+00 -5.08320808e-01 4.86728966e-01
2.89280027e-01 -5.76697439e-02 1.41337562e+00 -6.14672065e-01
7.35380709e-01 1.43721986e+00 8.05295467e-01 5.90986609e-01
-1.45386863e+00 -5.39234102e-01 2.15917170e-01 -1.64285917e-02
-9.60974276e-01 -6.12519145e-01 1.96106657e-01 -3.35355550e-02
1.68755031e+00 -1.04909264e-01 9.54239517e-02 1.21149850e+00
6.13666296e-01 8.80360901e-01 1.21020114e+00 -7.36085236e-01
-1.24740854e-01 -2.67710626e-01 4.59625453e-01 8.09580147e-01
-1.10014312e-01 3.79187465e-01 -7.57922292e-01 2.78377295e-01
4.92410034e-01 -6.99711561e-01 3.21316011e-02 3.13007027e-01
-9.39770818e-01 9.91448045e-01 5.29713869e-01 7.16034532e-01
-2.45528057e-01 2.14756176e-01 4.75720942e-01 2.15008676e-01
4.89669204e-01 4.97811168e-01 -9.00551200e-01 -5.03103375e-01
-9.18299019e-01 7.35911578e-02 7.33280420e-01 1.23052585e+00
7.33257294e-01 -6.59245774e-02 -5.52257895e-01 1.02296507e+00
1.59542292e-01 4.54044163e-01 4.76085663e-01 -7.24303424e-01
7.46347368e-01 3.30847725e-02 -2.95797616e-01 -3.86705786e-01
-7.18146741e-01 -3.03874254e-01 -7.04389691e-01 -3.37155610e-01
7.61726022e-01 -1.60651267e-01 -1.01863384e+00 2.07651258e+00
-1.31146912e-03 -2.80561537e-01 1.73452884e-01 5.22764504e-01
7.41823852e-01 8.50530267e-01 4.76556003e-01 -1.83960453e-01
1.49660325e+00 -1.11591733e+00 -7.59346545e-01 -4.54427689e-01
9.95417535e-01 -8.53365242e-01 9.06829655e-01 1.64646372e-01
-1.17588711e+00 -6.51045740e-01 -6.71528637e-01 -4.15129423e-01
-1.93123549e-01 2.28940159e-01 7.49940634e-01 7.32483447e-01
-1.27350783e+00 7.12347925e-01 -9.19952214e-01 -4.65080231e-01
2.42231771e-01 1.99176028e-01 -4.65507090e-01 -2.89669424e-01
-1.30374444e+00 1.21474349e+00 4.23824817e-01 -1.26176730e-01
-8.41128528e-01 -6.53346360e-01 -1.05106723e+00 2.05840170e-01
1.06383048e-01 -1.09352484e-01 1.55119896e+00 -3.28338087e-01
-1.70885146e+00 6.39800847e-01 -5.82168818e-01 -7.09508359e-01
-5.96052706e-02 -4.81599629e-01 -4.73591268e-01 -9.05501246e-02
3.25405896e-01 8.78878355e-01 6.76242262e-02 -6.82166040e-01
-7.88725555e-01 -1.73994094e-01 2.14932230e-03 -1.47045135e-01
-7.05723315e-02 3.30776125e-01 -2.40753412e-01 -3.55004460e-01
2.09328737e-02 -7.40742505e-01 -3.03300887e-01 -8.92101049e-01
-5.68570435e-01 -7.89047241e-01 2.99421012e-01 -8.33240092e-01
1.09564757e+00 -2.11245179e+00 -3.56037259e-01 -8.91408697e-02
-3.70275617e-01 3.71015251e-01 -6.70215249e-01 5.22714853e-01
-9.36092064e-02 2.95915037e-01 6.62286673e-03 -3.28997344e-01
1.23230703e-01 4.17997152e-01 -2.16495901e-01 1.51034564e-01
6.74623668e-01 1.10829961e+00 -9.76723373e-01 -7.01990068e-01
2.96984464e-02 3.05322170e-01 -3.77824306e-01 2.31481135e-01
-1.01359285e-01 3.08837473e-01 -1.91164926e-01 6.78414285e-01
5.77897668e-01 5.07205203e-02 6.12217724e-01 1.27730981e-01
-3.09270084e-01 1.11361575e+00 -4.26045150e-01 1.94389784e+00
-9.45453405e-01 6.45223916e-01 1.74867138e-02 -9.27430987e-01
1.09022653e+00 5.02734184e-01 -8.32302496e-02 -1.12237120e+00
-1.59303680e-01 1.37708053e-01 3.51986766e-01 -1.00839637e-01
7.28126705e-01 -1.72050476e-01 -4.48568165e-01 2.87055492e-01
5.45427978e-01 2.17329875e-01 5.18453419e-01 1.96462080e-01
1.42601991e+00 3.93464118e-01 5.65632105e-01 -5.92290521e-01
2.29983628e-01 7.95568526e-03 6.13154054e-01 1.05205178e+00
-1.81881636e-01 3.74758571e-01 5.74919999e-01 -4.56940621e-01
-9.07335520e-01 -1.08884561e+00 -5.47716975e-01 1.48816800e+00
-3.64661455e-01 -6.52422547e-01 -6.37175322e-01 -7.88229346e-01
-3.10398430e-01 9.35535252e-01 -4.30599540e-01 2.11341470e-01
-1.24127471e+00 -6.75352275e-01 9.32820201e-01 1.01861000e+00
3.21835548e-01 -1.27209854e+00 -2.06800044e-01 7.56735206e-01
-2.07925305e-01 -1.32431400e+00 -3.27728778e-01 9.28556800e-01
-9.95121658e-01 -4.23560321e-01 -3.51170659e-01 -1.13063478e+00
1.69257507e-01 -1.62886485e-01 1.50965631e+00 -2.67988890e-01
1.29034191e-01 -3.47237051e-01 -6.23926759e-01 -1.44589558e-01
-6.09754920e-01 6.36480570e-01 -3.14831704e-01 -7.63232231e-01
5.14370024e-01 -3.31914425e-01 8.16333946e-03 1.70451581e-01
-2.78994709e-01 -1.30893156e-01 9.16265428e-01 1.14499092e+00
4.88872856e-01 -4.25772190e-01 7.08465040e-01 -9.82310653e-01
4.86851960e-01 -2.42057666e-01 -4.86474365e-01 3.13166648e-01
-4.98039663e-01 5.04000843e-01 7.38771558e-01 -8.27843919e-02
-1.47504783e+00 1.00712530e-01 -4.30646271e-01 4.50538814e-01
-5.09704232e-01 5.71046889e-01 -2.11370125e-01 4.94576603e-01
6.57186270e-01 -4.41344790e-02 -4.78702217e-01 -8.08925450e-01
5.84841371e-01 5.63912392e-01 6.02278113e-01 -1.07723784e+00
3.55807990e-01 -1.74019277e-01 -2.40885541e-01 -5.94471753e-01
-1.10008764e+00 -5.49301565e-01 -7.79409647e-01 2.51655281e-01
9.11289811e-01 -9.91189718e-01 -4.72552627e-01 1.13105543e-01
-1.48104930e+00 -5.93293607e-01 -1.42784640e-01 3.19637775e-01
-4.61207509e-01 6.09800406e-02 -1.14584994e+00 -6.54099524e-01
-3.67210865e-01 -1.15668273e+00 1.21770501e+00 1.43933177e-01
-3.65310669e-01 -8.91970336e-01 3.27060223e-01 3.67421545e-02
2.38502845e-01 -1.41253665e-01 9.68711257e-01 -8.18737805e-01
-4.88862216e-01 1.47502542e-01 -2.09419653e-01 2.00187638e-01
-8.95876437e-02 -2.20931396e-01 -1.06153095e+00 -1.70372039e-01
-5.53607523e-01 -4.76782233e-01 9.99053955e-01 4.93384778e-01
7.12393939e-01 -6.31856248e-02 -4.56928521e-01 5.43489099e-01
1.33574259e+00 2.49712467e-02 5.47531545e-01 3.98816675e-01
4.18749183e-01 6.07681632e-01 6.18646085e-01 1.76889729e-02
3.79676491e-01 7.05407262e-01 7.97931924e-02 1.15885459e-01
-2.84592777e-01 -4.10016388e-01 7.27916300e-01 1.07152689e+00
3.41252945e-02 -4.86427605e-01 -1.03490782e+00 9.27157938e-01
-1.70155811e+00 -7.56174922e-01 -4.82396185e-01 1.75156248e+00
1.18579209e+00 2.29938418e-01 -1.32935390e-01 -2.45106712e-01
7.62416899e-01 2.83189267e-01 1.15962602e-01 -1.24868572e+00
-2.58915931e-01 9.06080961e-01 5.68763077e-01 4.73920554e-01
-1.23546231e+00 1.50472426e+00 6.94299841e+00 1.05333984e+00
-7.79621065e-01 4.58038807e-01 5.69335341e-01 7.43155554e-02
-2.07141824e-02 2.68458948e-02 -1.33211124e+00 2.05148399e-01
1.73301327e+00 2.24186495e-01 2.78071731e-01 8.61755610e-01
-1.81222096e-01 -2.75155276e-01 -1.22704959e+00 2.16806874e-01
-2.93269962e-01 -1.24108672e+00 -4.90520895e-01 8.42671320e-02
6.56107187e-01 6.02855146e-01 -2.35705316e-01 7.83368826e-01
1.10509169e+00 -1.22334039e+00 7.68783927e-01 1.43866003e-01
8.12866032e-01 -7.81238496e-01 9.52241659e-01 3.08288366e-01
-1.31764853e+00 1.87433958e-01 -6.12301052e-01 -1.35616452e-01
5.21693528e-01 6.69095278e-01 -1.16696572e+00 6.96619391e-01
8.00077379e-01 6.33676589e-01 -4.78406221e-01 6.78215861e-01
-8.43017519e-01 1.15204775e+00 -3.48596632e-01 -2.11140394e-01
5.32805324e-01 2.25912437e-01 1.70856863e-01 1.81968558e+00
2.99132406e-03 -1.42457470e-01 4.58758362e-02 5.47543168e-01
-8.56839791e-02 2.55191088e-01 -5.40017664e-01 2.40852565e-01
5.12163520e-01 1.20823133e+00 -4.79653627e-01 -2.34210610e-01
-4.49399650e-01 5.45492113e-01 1.16494584e+00 1.03456475e-01
-7.46925890e-01 -2.40389258e-01 5.34955025e-01 -4.20339584e-01
7.50560164e-01 -2.11801276e-01 -3.35701793e-01 -7.00856507e-01
-1.40397936e-01 -7.85600781e-01 4.93642688e-01 -4.65347797e-01
-1.41386414e+00 9.62018013e-01 -1.64383188e-01 -6.10905111e-01
-3.92698020e-01 -8.57003272e-01 -5.65951765e-01 1.09992266e+00
-1.63142836e+00 -1.38382936e+00 3.19249392e-01 1.53281912e-01
4.01139110e-01 -4.45625512e-04 1.30433583e+00 9.08426046e-02
-4.48144108e-01 9.22422051e-01 2.55656719e-01 4.62531447e-01
9.07722116e-01 -1.53141797e+00 8.94450307e-01 1.13325894e+00
5.02556503e-01 7.55495608e-01 2.82646865e-01 -5.93343616e-01
-8.09752226e-01 -1.11598980e+00 1.69954729e+00 -5.21203697e-01
6.57343805e-01 -4.81010139e-01 -6.97734416e-01 1.03138244e+00
5.29285967e-01 4.95456010e-02 4.76287574e-01 1.03966856e+00
-5.20972550e-01 -9.48063731e-02 -8.52049589e-01 1.67812690e-01
1.05963099e+00 -5.68512917e-01 -8.11086357e-01 2.89273590e-01
8.22360754e-01 -5.31841338e-01 -1.05734348e+00 2.94449449e-01
3.61981809e-01 -9.68992472e-01 7.19141543e-01 -6.22148156e-01
6.52409434e-01 1.74696833e-01 -1.69458434e-01 -1.48350930e+00
-7.14085042e-01 -5.21355748e-01 2.72912890e-01 1.47428560e+00
8.38753104e-01 -4.14847493e-01 4.90912378e-01 1.26860186e-01
-6.77253485e-01 -6.14043951e-01 -1.28364289e+00 -1.02048612e+00
5.98450780e-01 -3.90352786e-01 6.30630851e-01 5.07682204e-01
2.53297508e-01 5.22185445e-01 -2.70743400e-01 1.06350042e-01
1.01611368e-01 1.39945492e-01 4.79124486e-01 -1.00540078e+00
-3.38206708e-01 -2.08554760e-01 -1.70670405e-01 -1.32322133e+00
5.06339908e-01 -1.22095573e+00 6.46427572e-01 -1.35861635e+00
2.50218004e-01 -7.22331583e-01 -2.76173443e-01 8.26490939e-01
-2.04848200e-01 1.56419292e-01 2.07574502e-01 -1.54046670e-01
-6.86976314e-01 2.15878218e-01 1.24763715e+00 8.23120624e-02
6.30089687e-03 -3.81979704e-01 -5.30565500e-01 3.72986466e-01
9.58589315e-01 -7.92394876e-01 1.29191145e-01 -8.81552637e-01
-4.44837660e-03 1.77885011e-01 -1.86986864e-01 -8.93200517e-01
6.21177666e-02 1.58627391e-01 2.35356227e-01 -6.39318824e-01
1.88240960e-01 -8.68970081e-02 -5.11111379e-01 2.85606176e-01
-4.74620730e-01 3.28460276e-01 4.12184238e-01 3.64192784e-01
-2.90419906e-01 -5.22221863e-01 6.27537310e-01 -1.98363408e-01
-6.14999592e-01 -5.16405962e-02 -5.38964987e-01 5.18896878e-01
4.08984452e-01 1.86142221e-01 -6.08174562e-01 -2.35571750e-02
-5.88721454e-01 -1.71735231e-02 4.52985726e-02 2.51660615e-01
-3.07903090e-03 -9.50564742e-01 -7.71459103e-01 2.11630329e-01
-3.88334021e-02 1.14059769e-01 1.28691699e-02 8.38668168e-01
-4.12725925e-01 8.72049689e-01 -1.33266941e-01 -5.90078712e-01
-1.04603338e+00 5.72962984e-02 4.78582643e-02 -9.06172872e-01
-5.61818242e-01 1.24862742e+00 -1.75106619e-02 -6.83811367e-01
-4.08719964e-02 -5.47370374e-01 -7.33903237e-03 -1.52335897e-01
3.87011975e-01 2.99367886e-02 3.28847617e-01 -7.47536898e-01
-5.75886846e-01 1.62043855e-01 -4.03484166e-01 -2.20666930e-01
1.53630877e+00 2.62914389e-01 -7.58692622e-02 2.14461938e-01
1.13486969e+00 1.02431893e-01 -1.12232518e+00 -2.39661559e-01
4.99074906e-01 1.37222977e-02 6.72963448e-03 -1.19760907e+00
-7.14096904e-01 1.06317711e+00 1.39749333e-01 -1.12548750e-02
7.98194289e-01 3.91303688e-01 1.05477464e+00 4.04873669e-01
4.69255328e-01 -1.03539908e+00 -1.93709135e-01 1.23516893e+00
4.31792587e-01 -1.17459178e+00 -2.41460055e-01 -3.29518855e-01
-5.11985838e-01 1.12072825e+00 6.01873934e-01 -1.58232808e-01
3.69981140e-01 6.17845595e-01 1.56484261e-01 8.19843411e-02
-9.73111272e-01 -2.49303728e-01 1.01177119e-01 7.13405848e-01
1.04345345e+00 4.95271981e-01 -4.66847688e-01 7.71558404e-01
-6.24395072e-01 -5.40120363e-01 2.99283564e-01 8.21351469e-01
-4.01001900e-01 -1.66236234e+00 -1.72669664e-01 2.71717042e-01
-7.29676664e-01 -5.99989831e-01 1.16641924e-03 1.18588829e+00
-1.11151775e-02 1.16757429e+00 2.33266145e-01 -1.68618992e-01
8.26681778e-02 4.04281706e-01 6.51108205e-01 -7.38549829e-01
-9.52418447e-01 1.59133285e-01 7.74117291e-01 -7.18705893e-01
-4.19347286e-01 -8.76387298e-01 -1.26031804e+00 -1.87723979e-01
-6.64326251e-01 2.98057407e-01 6.22824132e-01 1.08616889e+00
1.56387165e-01 5.42869687e-01 2.38618255e-01 -4.39745933e-01
-6.93086565e-01 -1.41500652e+00 -4.14105833e-01 2.25529030e-01
2.08685417e-02 -4.02719706e-01 1.01924650e-01 -5.39433323e-02]
|
[10.341704368591309, 9.727469444274902]
|
05edb334-8fd1-4291-8d11-2954f350a934
|
dialogue-act-recognition-using-reweighted
| null | null |
https://aclanthology.org/W12-1616
|
https://aclanthology.org/W12-1616.pdf
|
Dialogue Act Recognition using Reweighted Speaker Adaptation
| null |
['Louis-Philippe Morency', 'Congkai Sun']
|
2012-07-01
| null | null | null |
ws-2012-7
|
['dialogue-understanding']
|
['natural-language-processing']
|
[-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01]
|
[-7.429410934448242, 3.570490598678589]
|
66ff8435-7bdb-4804-bdb8-c73eb5daeae3
|
conformal-predictors-for-compound-activity
|
1603.04506
| null |
http://arxiv.org/abs/1603.04506v1
|
http://arxiv.org/pdf/1603.04506v1.pdf
|
Conformal Predictors for Compound Activity Prediction
|
The paper presents an application of Conformal Predictors to a
chemoinformatics problem of identifying activities of chemical compounds. The
paper addresses some specific challenges of this domain: a large number of
compounds (training examples), high-dimensionality of feature space, sparseness
and a strong class imbalance. A variant of conformal predictors called
Inductive Mondrian Conformal Predictor is applied to deal with these
challenges. Results are presented for several non-conformity measures (NCM)
extracted from underlying algorithms and different kernels. A number of
performance measures are used in order to demonstrate the flexibility of
Inductive Mondrian Conformal Predictors in dealing with such a complex set of
data.
Keywords: Conformal Prediction, Confidence Estimation, Chemoinformatics,
Non-Conformity Measure.
|
['Alexander Gammerman', 'Paolo Toccacheli', 'Ilia Nouretdinov']
|
2016-03-14
| null | null | null | null |
['activity-prediction', 'activity-prediction']
|
['computer-vision', 'time-series']
|
[ 4.18680280e-01 -2.56225377e-01 -1.50150880e-01 -4.31452543e-01
-6.85069621e-01 -7.94867039e-01 9.07989502e-01 4.98614788e-01
-5.58413416e-02 1.19006395e+00 -4.69398238e-02 -4.65655744e-01
-7.06221163e-01 -5.71254969e-01 -3.04519206e-01 -9.86371577e-01
-4.80570227e-01 6.77280486e-01 1.71526119e-01 -3.05352602e-02
7.72505462e-01 7.84433722e-01 -1.28107536e+00 4.07278091e-01
1.00595737e+00 5.37388623e-01 1.94439843e-01 6.81110203e-01
1.40853330e-01 3.46667379e-01 -1.44326121e-01 -4.39731747e-01
1.11275151e-01 -4.12785597e-02 -6.91184878e-01 -8.08704376e-01
5.87605417e-01 5.04740238e-01 3.95924956e-01 9.57647026e-01
5.05204201e-01 1.27456278e-01 1.71378732e+00 -1.25393295e+00
-4.89889234e-01 -7.83377364e-02 -4.69136894e-01 3.51053566e-01
5.02150476e-01 -3.63032788e-01 7.15554535e-01 -1.37291169e+00
3.85271758e-01 1.02417445e+00 1.00463033e+00 1.60189092e-01
-1.33932161e+00 -7.32702672e-01 -3.46282423e-01 4.80144769e-01
-1.42335725e+00 2.14285683e-02 2.19229162e-01 -9.41325128e-01
9.36709881e-01 5.61108232e-01 2.88683385e-01 1.07975721e+00
1.05145991e+00 2.89951451e-02 1.54000747e+00 -4.47370440e-01
8.21573019e-01 5.26696146e-01 3.33189905e-01 2.93755978e-01
2.31292933e-01 5.10459542e-01 -4.38622773e-01 -7.56248593e-01
2.81819999e-01 4.74351877e-03 -9.80380848e-02 -5.62775195e-01
-7.12775588e-01 1.24609935e+00 2.44836837e-01 3.10470700e-01
-2.57139593e-01 -4.77184296e-01 4.88297194e-02 1.25502750e-01
5.97261429e-01 7.72606194e-01 -8.48962724e-01 1.16491608e-01
-5.71879983e-01 2.21974805e-01 8.98564398e-01 1.28527200e+00
4.69328225e-01 -2.42920265e-01 -1.91037610e-01 4.92533863e-01
1.80320904e-01 5.08262336e-01 3.51154596e-01 -6.24219552e-02
2.15319976e-01 2.47795373e-01 -2.63069030e-02 -6.95601463e-01
-7.39909768e-01 -8.41585100e-02 -7.36883521e-01 3.23327690e-01
4.18271035e-01 -1.50570571e-02 -8.24346244e-01 1.19899654e+00
5.32733023e-01 8.17291960e-02 2.38218546e-01 5.30768216e-01
7.51568317e-01 7.65836298e-01 7.55939245e-01 -2.31285036e-01
1.41785228e+00 -6.18181229e-01 -2.04753876e-01 3.30499053e-01
6.40676618e-01 -1.15474772e+00 7.93979645e-01 5.58836341e-01
-7.07219779e-01 -2.41969690e-01 -1.11330879e+00 1.12046242e-01
-1.01012504e+00 1.05953246e-01 8.73044670e-01 7.43934512e-01
-4.33605492e-01 1.04697740e+00 -2.91574538e-01 -1.83668643e-01
4.64424610e-01 6.70836151e-01 -6.62453771e-01 -9.10455734e-02
-9.01246786e-01 1.29821527e+00 1.15668990e-01 -6.69847488e-01
-6.90950215e-01 -1.37177348e+00 -4.26733971e-01 7.45275468e-02
-3.40096712e-01 -1.59970537e-01 7.42343068e-01 -5.72974384e-01
-1.14593065e+00 4.32132185e-01 2.18054503e-01 3.00058592e-02
3.30633521e-01 -8.07538331e-02 -5.81694186e-01 -2.60492504e-01
2.63801310e-02 2.54690051e-01 3.80500138e-01 -5.47639787e-01
-4.80424017e-01 -6.06370151e-01 -4.85542625e-01 2.51374859e-02
-3.39800954e-01 8.34703967e-02 5.38221478e-01 -7.25108087e-01
-2.96076357e-01 -1.09015143e+00 -3.71611923e-01 -5.57394028e-01
-3.22802871e-01 -3.29246372e-01 7.41359055e-01 -3.11972827e-01
6.48057282e-01 -1.72600687e+00 2.93139845e-01 5.91193020e-01
-3.34662721e-02 1.27645209e-01 -2.34858841e-01 7.39913344e-01
-8.20221186e-01 -1.88104466e-01 -4.57656980e-01 4.43409026e-01
-3.78332883e-01 -2.78130561e-01 -3.21026176e-01 7.29013503e-01
3.92977148e-01 6.30862296e-01 -7.72642910e-01 -4.01885629e-01
-2.54434943e-02 9.59419549e-01 -2.29177713e-01 1.20942824e-01
-3.17361020e-02 5.08483410e-01 -5.48466802e-01 8.35588455e-01
9.19817150e-01 -2.93819100e-01 1.62997410e-01 -1.59364417e-01
-3.05526167e-01 -1.64775714e-01 -9.68614161e-01 1.21119905e+00
-1.07484140e-01 2.84295708e-01 -7.90635824e-01 -7.63749778e-01
8.82077277e-01 4.12547797e-01 4.99154717e-01 -6.01720095e-01
-2.41238009e-02 1.33262798e-01 4.90597971e-02 -1.37930065e-01
-2.28983928e-02 -2.69855291e-01 4.06831831e-01 2.40405411e-01
4.97807324e-01 -6.94031939e-02 1.31700292e-01 -8.09105635e-02
1.04640329e+00 1.50402874e-01 7.65386701e-01 -1.20405853e+00
8.16741109e-01 1.02865733e-01 3.40354800e-01 9.58170965e-02
1.12414032e-01 3.93950850e-01 3.72739196e-01 -9.75748777e-01
-1.37778687e+00 -9.80697870e-01 -1.08353603e+00 1.03404975e+00
-3.26064646e-01 -5.85935354e-01 -8.47865820e-01 -9.08977747e-01
1.95954517e-01 5.83633184e-01 -8.86844397e-01 -5.56847565e-02
-1.06168270e-01 -1.40376139e+00 6.93636164e-02 4.38778192e-01
-2.46982574e-01 -5.69307327e-01 -2.08763793e-01 9.74660367e-02
6.33208573e-01 -6.72478020e-01 -2.42496595e-01 7.36368716e-01
-9.26274538e-01 -1.68132401e+00 -6.79462552e-01 -6.94539607e-01
6.96363151e-01 3.50872763e-02 1.10463774e+00 -5.97831845e-01
-7.27806389e-01 8.63007084e-02 -1.15517840e-01 -1.01000869e+00
-3.72916967e-01 -1.07075274e-01 1.14718519e-01 -3.45313072e-01
9.34889495e-01 -7.41451859e-01 -5.25453866e-01 5.13251841e-01
-6.41961336e-01 -2.98952132e-01 8.19560945e-01 8.86592269e-01
7.54283965e-01 -7.38952011e-02 6.78812981e-01 -1.35464168e+00
7.07238317e-01 -8.23080182e-01 -9.79415655e-01 5.15331507e-01
-1.08028781e+00 1.43261999e-01 7.19593108e-01 -8.41808677e-01
-6.81054413e-01 3.17679465e-01 2.03246444e-01 1.40875131e-01
-1.54684246e-01 2.99977094e-01 2.14772187e-02 -6.29943430e-01
1.31051421e+00 -1.62392519e-02 -2.82859802e-01 -4.70119625e-01
2.31978763e-02 5.17690778e-01 3.58176348e-03 -4.69940931e-01
6.76578701e-01 4.23175506e-02 9.14890885e-01 -7.14113712e-01
-5.97702265e-01 -5.50101101e-01 -8.65449965e-01 2.67921120e-01
6.74520493e-01 -8.58521342e-01 -8.54398668e-01 -1.50811940e-01
-8.92520726e-01 1.04787163e-01 -1.77447170e-01 7.96178102e-01
-6.08144045e-01 1.18587196e-01 -3.49206328e-01 -5.99007607e-01
-7.20484257e-01 -1.05086768e+00 8.03278804e-01 1.30233526e-01
-3.80334258e-01 -1.20999944e+00 7.33062208e-01 -1.55230701e-01
4.44721162e-01 4.57746416e-01 1.33183336e+00 -9.95214581e-01
-1.23185953e-02 -3.69012982e-01 6.05393387e-02 -2.17532665e-01
1.68324307e-01 9.76028368e-02 -1.41035509e+00 -2.13922739e-01
-7.78831318e-02 -1.48461878e-01 4.84529108e-01 5.79276502e-01
1.07543099e+00 -1.95135385e-01 -5.61790884e-01 5.80207288e-01
1.37710071e+00 4.62554604e-01 5.90750098e-01 1.43885255e-01
4.07826304e-01 6.42228544e-01 8.26707900e-01 4.10412997e-01
-1.01283275e-01 6.39774501e-01 1.30513787e-01 -1.41748518e-01
3.81192625e-01 -6.28438592e-02 1.54931853e-02 4.51009840e-01
-4.21773553e-01 2.30297148e-01 -9.29236233e-01 -1.78349093e-01
-1.53291464e+00 -7.23692060e-01 -6.48559868e-01 2.30995631e+00
9.89674389e-01 -5.29859722e-01 1.58742875e-01 1.65347949e-01
5.60530365e-01 -7.72753596e-01 -4.48785871e-01 -6.18569613e-01
-3.20546985e-01 6.51425004e-01 6.70784652e-01 4.78115976e-01
-1.47393441e+00 6.77243054e-01 7.66262436e+00 1.17444658e+00
-9.04818594e-01 -9.12986137e-03 8.14515471e-01 4.18647557e-01
-1.62651554e-01 1.30040333e-01 -9.87843573e-01 6.53840363e-01
1.37352931e+00 -2.90593028e-01 7.60475099e-02 1.07103598e+00
-4.63735282e-01 -1.14561722e-01 -1.30145121e+00 8.21113050e-01
1.06216900e-01 -1.51128829e+00 3.15965191e-02 2.31035471e-01
9.50617313e-01 5.35049848e-03 2.72723734e-01 -1.32408813e-01
3.24048668e-01 -1.44443130e+00 -2.79612411e-02 6.42247856e-01
9.53183949e-01 -1.07846785e+00 6.15120828e-01 5.56995980e-02
-1.02687681e+00 -3.22812870e-02 -8.55030417e-01 3.57253104e-01
-6.39251709e-01 7.80620813e-01 -1.15239060e+00 6.24951422e-01
5.96392274e-01 8.00895452e-01 -9.31091309e-01 1.35122013e+00
2.14822531e-01 2.02729389e-01 -7.58418515e-02 -4.05672878e-01
1.11750789e-01 -5.80488861e-01 1.60301179e-01 1.40579760e+00
6.67856872e-01 2.17557460e-01 -3.68159175e-01 6.74116194e-01
4.48744982e-01 7.02251494e-01 -9.92085934e-01 2.44529456e-01
3.20170283e-01 1.44925809e+00 -4.96646613e-01 -8.60867053e-02
-4.06935006e-01 4.27748024e-01 2.30791718e-01 1.84151635e-01
-6.40831947e-01 -1.01546571e-01 4.23359036e-01 1.30079031e-01
1.17812650e-02 7.45471567e-02 -2.94542432e-01 -7.65828669e-01
-5.65866768e-01 -7.41841733e-01 7.10115373e-01 -6.55326247e-01
-1.54394746e+00 5.49262762e-01 6.42209351e-02 -1.53402662e+00
2.73181707e-01 -1.18448794e+00 -7.52353311e-01 1.11462760e+00
-1.39993370e+00 -9.53054130e-01 -2.20983494e-02 6.53161585e-01
2.14602679e-01 -9.35648024e-01 1.66924036e+00 2.50987858e-01
-1.17000662e-01 3.54623079e-01 5.58471799e-01 -8.36292982e-01
1.08535218e+00 -1.68873990e+00 1.02259375e-01 2.49453094e-02
1.32433906e-01 5.22482395e-01 5.97620487e-01 -6.14194989e-01
-1.43011904e+00 -1.08127296e+00 8.10767353e-01 -9.32502508e-01
5.28564274e-01 -2.04086334e-01 -7.73183763e-01 1.66977465e-01
-5.42050228e-02 -4.65803035e-02 1.71015584e+00 3.57351243e-01
-4.83144373e-01 4.71105762e-02 -1.27045655e+00 2.90671557e-01
3.33555192e-01 -1.92379892e-01 -3.05284858e-01 1.06542826e+00
5.36582246e-02 -6.64648414e-02 -1.49432290e+00 5.75459063e-01
4.88680840e-01 -5.56950927e-01 1.17919278e+00 -1.24933171e+00
3.52805197e-01 -3.12103689e-01 -9.82820839e-02 -1.36556792e+00
-7.12567866e-01 -4.61138636e-01 -4.57155565e-03 5.98264813e-01
6.29174352e-01 -3.98332715e-01 8.93550277e-01 5.95286727e-01
-5.99818639e-02 -8.05951357e-01 -1.12555420e+00 -7.21015513e-01
6.90153301e-01 1.75643429e-01 5.36866784e-01 1.29201436e+00
5.64536214e-01 8.69328737e-01 -3.93993586e-01 4.20979746e-02
6.62571728e-01 4.41109501e-02 5.46750546e-01 -1.81562757e+00
-6.46486655e-02 -9.91198421e-02 -8.88888299e-01 7.97788873e-02
-7.97796547e-02 -8.83008957e-01 -5.52803576e-01 -6.19596124e-01
6.14332557e-01 -7.11071372e-01 -2.95251727e-01 2.92312264e-01
-1.19607836e-01 7.99791664e-02 -3.75243604e-01 7.01196399e-03
-1.75844327e-01 2.89097488e-01 1.00410736e+00 -4.15847927e-01
-6.06563464e-02 1.49792731e-01 -6.39950216e-01 5.82366943e-01
8.05579066e-01 -6.60741329e-01 -5.98774135e-01 5.78234255e-01
3.25674027e-01 -1.63456619e-01 -3.78653617e-03 -1.03590727e+00
3.36926162e-01 -4.43786621e-01 1.09664237e+00 -4.97817099e-01
3.30987483e-01 -9.60914075e-01 5.76115429e-01 4.33653712e-01
-2.04692017e-02 3.38013351e-01 3.67909163e-01 7.11965978e-01
-8.36461131e-03 -3.73413086e-01 9.39140558e-01 2.76911259e-01
-1.40615463e-01 4.89605784e-01 -2.46405274e-01 -2.44516164e-01
1.54389906e+00 -2.64655113e-01 -3.58392239e-01 2.78162390e-01
-7.20926046e-01 -1.50424600e-01 2.69577235e-01 1.18827954e-01
4.39790845e-01 -1.51466727e+00 -6.50029361e-01 3.68443102e-01
4.95399356e-01 -3.98143023e-01 -2.66577192e-02 9.99699712e-01
-4.44950432e-01 7.59916008e-01 -5.82743883e-01 -6.38178766e-01
-1.61117506e+00 7.36687839e-01 1.64790869e-01 -1.24878794e-01
-3.06294084e-01 6.62097692e-01 5.80853403e-01 -2.96526968e-01
1.39285503e-02 4.46417145e-02 -6.52468324e-01 6.49064779e-03
6.62617624e-01 4.06915069e-01 5.00451684e-01 -3.95552218e-01
-4.62672770e-01 9.23786521e-01 -1.53016359e-01 4.54627037e-01
1.67044628e+00 8.51681888e-01 -1.19077377e-02 1.26863316e-01
1.01687431e+00 -1.87534064e-01 -1.11469066e+00 9.41599905e-02
3.74902964e-01 -5.32711387e-01 -2.25113124e-01 -1.01819742e+00
-3.30048680e-01 9.24263835e-01 1.03498256e+00 -1.60043374e-01
6.85248673e-01 -1.82244360e-01 -2.02750131e-01 6.09350741e-01
8.48712549e-02 -1.26572692e+00 -9.02049169e-02 4.54278618e-01
1.11197400e+00 -1.25328994e+00 5.69944263e-01 -6.69268250e-01
-6.38573647e-01 1.53439558e+00 2.23019406e-01 2.87949350e-02
1.15250421e+00 3.99129748e-01 -3.54477614e-01 -3.25622320e-01
-6.30589306e-01 3.09781522e-01 6.49397790e-01 9.93886650e-01
9.01213944e-01 2.28566289e-01 -5.35022914e-01 4.36443657e-01
-1.37308806e-01 -4.39112008e-01 2.43770912e-01 9.79716480e-01
-3.12907219e-01 -1.44922483e+00 -4.20616537e-01 4.67171758e-01
-5.37713110e-01 -4.29305375e-01 -7.08175421e-01 6.91955030e-01
-1.21114679e-01 6.11346006e-01 -3.08427691e-01 -2.55819827e-01
2.34409168e-01 1.54623706e-02 6.08747602e-01 -3.45534593e-01
-3.88277888e-01 4.64161783e-02 -9.63291302e-02 -2.20256537e-01
-3.23956609e-01 -5.63634217e-01 -8.98183107e-01 -4.35202807e-01
-5.66390753e-01 7.06516504e-01 9.78205323e-01 6.20944738e-01
3.05436343e-01 1.94356233e-01 6.16741776e-01 -6.25652432e-01
-6.54724181e-01 -1.13625538e+00 -6.80295050e-01 3.30626070e-01
-1.83410361e-01 -9.33342218e-01 -1.71787292e-01 2.05070436e-01]
|
[5.132468223571777, 5.586060047149658]
|
74f203fa-0ec3-48de-8f04-cfc41951b23f
|
dual-tasks-siamese-transformer-framework-for
|
2201.10953
| null |
https://arxiv.org/abs/2201.10953v2
|
https://arxiv.org/pdf/2201.10953v2.pdf
|
Dual-Tasks Siamese Transformer Framework for Building Damage Assessment
|
Accurate and fine-grained information about the extent of damage to buildings is essential for humanitarian relief and disaster response. However, as the most commonly used architecture in remote sensing interpretation tasks, Convolutional Neural Networks (CNNs) have limited ability to model the non-local relationship between pixels. Recently, Transformer architecture first proposed for modeling long-range dependency in natural language processing has shown promising results in computer vision tasks. Considering the frontier advances of Transformer architecture in the computer vision field, in this paper, we present the first attempt at designing a Transformer-based damage assessment architecture (DamFormer). In DamFormer, a siamese Transformer encoder is first constructed to extract non-local and representative deep features from input multitemporal image-pairs. Then, a multitemporal fusion module is designed to fuse information for downstream tasks. Finally, a lightweight dual-tasks decoder aggregates multi-level features for final prediction. To the best of our knowledge, it is the first time that such a deep Transformer-based network is proposed for multitemporal remote sensing interpretation tasks. The experimental results on the large-scale damage assessment dataset xBD demonstrate the potential of the Transformer-based architecture.
|
['Lars Bromley', 'Chen Wu', 'Xi Li', 'Sofia Vallecorsa', 'Edoardo Nemni', 'Hongruixuan Chen']
|
2022-01-26
| null | null | null | null |
['extracting-buildings-in-remote-sensing-images']
|
['miscellaneous']
|
[ 4.74554658e-01 -1.50937065e-01 3.06922883e-01 -5.84886134e-01
-1.05604744e+00 -4.83736657e-02 5.60568869e-01 3.17423493e-01
-5.01202226e-01 5.71934581e-01 5.21011293e-01 -2.92064011e-01
-2.21325070e-01 -1.06649339e+00 -5.01887798e-01 -7.03593612e-01
-2.84289032e-01 1.75077334e-01 9.94843617e-02 -4.50697631e-01
4.74976189e-02 6.50680125e-01 -1.55246270e+00 5.73223054e-01
8.71581614e-01 1.23902893e+00 4.82977659e-01 4.49633837e-01
2.82795846e-01 8.55528772e-01 -2.03732669e-01 -1.35108724e-01
1.96766734e-01 -1.61249220e-01 -8.94687057e-01 -4.08095531e-02
4.04551059e-01 -8.47839594e-01 -5.28982818e-01 8.01730037e-01
8.01828146e-01 -9.33211297e-02 7.56771028e-01 -5.80335855e-01
-6.93178833e-01 7.18042016e-01 -6.50998533e-01 6.22354388e-01
1.38991147e-01 9.25791115e-02 1.22996211e+00 -1.06666481e+00
1.49866328e-01 1.18693912e+00 7.80809104e-01 4.87489440e-03
-7.93836951e-01 -3.38186294e-01 2.08182961e-01 4.17708248e-01
-1.34584665e+00 -9.14064124e-02 9.03760374e-01 -3.55042756e-01
1.20634139e+00 1.74119651e-01 5.89487672e-01 1.00497639e+00
2.90251911e-01 9.58715498e-01 1.11846411e+00 -1.99635431e-01
5.87539934e-02 -6.69930696e-01 3.41521688e-02 7.47989595e-01
-2.09128067e-01 1.00806959e-01 -5.44129014e-01 1.62982896e-01
5.97074449e-01 2.30937526e-01 -2.19833925e-01 2.74271637e-01
-1.32190299e+00 7.47656226e-01 1.20662022e+00 5.65985978e-01
-7.73634851e-01 4.35912430e-01 3.61271203e-01 1.76206201e-01
8.28140855e-01 -1.35902287e-02 -2.40908608e-01 2.67408609e-01
-1.24401617e+00 3.77023034e-02 2.35292494e-01 3.12344372e-01
7.20612049e-01 9.81728286e-02 -3.07083398e-01 8.33482265e-01
3.06472987e-01 6.35478199e-01 2.18127206e-01 -5.16090572e-01
7.43154883e-01 6.04232252e-01 -2.44721204e-01 -1.06552494e+00
-7.45369375e-01 -7.83598661e-01 -1.40282774e+00 2.34836891e-01
1.20849289e-01 1.21529832e-01 -9.33738053e-01 1.46519935e+00
-5.59534542e-02 -8.49208757e-02 -1.55382365e-01 1.24159348e+00
6.29219472e-01 7.75303066e-01 2.39033848e-01 2.52082884e-01
1.35878897e+00 -5.85697651e-01 -4.33695853e-01 -4.56031889e-01
2.63513267e-01 -4.18848008e-01 9.96377289e-01 2.86922157e-01
-7.22505987e-01 -4.79364365e-01 -1.09775829e+00 -3.37228298e-01
-4.31723654e-01 2.46397018e-01 4.14582163e-01 1.98373184e-01
-1.00793850e+00 4.86849606e-01 -8.98775995e-01 -4.53676969e-01
7.19406843e-01 6.69115689e-03 -3.66883308e-01 -1.98646501e-01
-1.42117023e+00 1.05093956e+00 2.77047575e-01 8.21215630e-01
-1.21342337e+00 -6.32879972e-01 -8.94063890e-01 2.12144807e-01
4.58450690e-02 -6.91978991e-01 9.61417913e-01 -6.17996037e-01
-8.80560577e-01 7.37564862e-01 7.02565834e-02 -5.85012972e-01
3.50165814e-01 -2.94405401e-01 -1.14618719e-01 1.83139995e-01
2.13121146e-01 6.47795856e-01 8.04015696e-01 -1.12986267e+00
-6.81708753e-01 -6.68760777e-01 1.60482571e-01 3.27148348e-01
-5.16161680e-01 2.63387952e-02 9.87963602e-02 -9.29332554e-01
3.61117035e-01 -5.19984722e-01 -3.66669953e-01 1.71963468e-01
-3.29156607e-01 -3.73258430e-04 6.61772192e-01 -1.01557779e+00
1.32163513e+00 -2.05673695e+00 2.00507760e-01 1.88963264e-01
2.58385897e-01 2.09114462e-01 -1.58581614e-01 6.28179908e-01
-9.56522599e-02 4.17131931e-02 -9.36744928e-01 -3.67091626e-01
7.29076937e-02 3.74674976e-01 -5.41523218e-01 4.91875321e-01
3.06473434e-01 1.04592454e+00 -8.45393062e-01 -3.17433238e-01
2.66995192e-01 6.03171110e-01 -2.72156805e-01 1.23128362e-01
4.72954176e-02 4.10435915e-01 -5.76592565e-01 9.16739821e-01
6.54251158e-01 4.16378630e-03 -2.38629192e-01 -5.27146041e-01
-2.40344509e-01 8.13569948e-02 -5.70089221e-01 1.90757108e+00
-5.76431155e-01 4.72634912e-01 2.18248274e-02 -1.18907058e+00
8.39836895e-01 2.27334589e-01 5.75978100e-01 -9.58984256e-01
-3.37537751e-02 4.40037042e-01 -3.70832145e-01 -4.65218961e-01
5.99485755e-01 -4.01683837e-01 -4.57837731e-01 3.83666009e-01
-2.50136405e-01 -2.26516232e-01 -1.82697475e-01 -1.34921312e-01
1.13651824e+00 2.71716435e-02 2.68234700e-01 -8.64023790e-02
5.44668019e-01 -7.92046718e-04 3.05348814e-01 5.40247381e-01
-7.95513019e-02 9.79709446e-01 -2.07732469e-02 -9.21223402e-01
-9.70467806e-01 -1.09978271e+00 -1.11918449e-01 1.06094599e+00
-2.64326513e-01 -8.45775660e-03 -5.73693216e-01 -4.56760675e-01
-1.28424063e-01 4.68050092e-01 -5.44091046e-01 -9.63552892e-02
-6.56118691e-01 -1.12904227e+00 1.04171669e+00 7.40038753e-01
1.04560137e+00 -1.01474547e+00 -9.49827790e-01 4.33659047e-01
-6.00902975e-01 -9.02483940e-01 -2.36762449e-01 2.52134562e-01
-7.94180393e-01 -8.18560362e-01 -7.54680037e-01 -5.86970031e-01
3.19816589e-01 2.96742022e-01 7.99073696e-01 -9.40308720e-03
-3.27369362e-01 4.31204244e-04 -4.24723089e-01 -1.52863696e-01
8.60426128e-02 2.15004534e-01 -3.24626654e-01 1.11980453e-01
-1.09965391e-02 -8.46537590e-01 -7.39498734e-01 1.22743003e-01
-1.17442715e+00 9.95927956e-04 9.70196366e-01 1.00012052e+00
5.23132086e-01 2.30271488e-01 4.80361044e-01 -1.62752017e-01
5.74522138e-01 -4.79441494e-01 -1.60543635e-01 3.94297272e-01
-3.26283991e-01 1.27264619e-01 5.22859693e-01 2.53094107e-01
-1.25723374e+00 1.08323440e-01 -5.07577419e-01 -7.54269809e-02
-3.16045403e-01 1.21253002e+00 -1.18620291e-01 3.36810231e-01
6.22631788e-01 3.72817159e-01 -4.82060164e-01 -7.73203969e-01
2.30204269e-01 8.53866577e-01 6.80029273e-01 -4.59373444e-01
7.05566406e-01 6.90962195e-01 1.68212757e-01 -8.40849042e-01
-1.11600983e+00 -3.67477864e-01 -8.74452889e-01 -2.05344826e-01
1.23796523e+00 -1.15410519e+00 -2.76973009e-01 1.10928607e+00
-1.18617201e+00 -3.06362629e-01 -1.82589307e-01 3.57077658e-01
-3.90862644e-01 3.58189046e-01 -4.89469558e-01 -6.67665541e-01
-6.18885934e-01 -1.08158028e+00 1.53381467e+00 -2.20496431e-01
2.63195574e-01 -8.78638983e-01 -8.21354836e-02 6.07500315e-01
6.79780602e-01 4.32244927e-01 6.67842567e-01 -1.17874146e-02
-5.89844644e-01 -4.40977924e-02 -4.69456941e-01 4.20107692e-01
9.03326795e-02 -4.86751348e-01 -1.07733393e+00 -2.39545092e-01
3.91522422e-03 -4.01026517e-01 1.37424636e+00 3.64933640e-01
1.21952331e+00 -1.26193702e-01 -5.73319830e-02 6.68042958e-01
1.46673977e+00 -2.75339007e-01 8.13618422e-01 3.90430182e-01
9.30468917e-01 4.46835846e-01 6.48365736e-01 6.44209802e-01
8.12189043e-01 5.68448067e-01 8.21432352e-01 -3.95808011e-01
-1.40766218e-01 -2.04122841e-01 4.89584446e-01 5.03020525e-01
-1.11983836e-01 -4.06120181e-01 -1.39946532e+00 8.38015199e-01
-1.74227417e+00 -1.19898736e+00 -2.74790972e-01 1.75882828e+00
4.83788043e-01 -1.98238403e-01 -4.09333497e-01 3.49817038e-01
3.36282670e-01 5.89922845e-01 -3.94525737e-01 -1.02380447e-01
-4.35228825e-01 3.27044755e-01 5.68780780e-01 5.41187584e-01
-1.48286104e+00 8.43100488e-01 5.28934956e+00 9.75202084e-01
-1.03958046e+00 2.98378468e-01 5.97250819e-01 -4.77750931e-04
-4.04000491e-01 -1.43104434e-01 -3.32482129e-01 6.16770834e-02
6.25439465e-01 4.06673372e-01 5.41638993e-02 1.97380811e-01
6.15230739e-01 -2.53101468e-01 -6.92014754e-01 7.87929595e-01
4.26419415e-02 -9.36318159e-01 1.25853792e-01 7.88198337e-02
4.60144281e-01 4.89582956e-01 2.26754054e-01 -1.44806400e-01
7.99681917e-02 -1.14010227e+00 1.03277731e+00 7.79584229e-01
6.37719393e-01 -8.04863632e-01 8.77324164e-01 4.98678654e-01
-1.43658078e+00 -4.58742172e-01 -3.16096157e-01 -2.02995762e-01
5.26691556e-01 9.64143634e-01 -6.18163526e-01 8.84349167e-01
9.89673316e-01 1.07055509e+00 -6.57412231e-01 7.82688856e-01
-3.87345344e-01 6.85923815e-01 -3.92764658e-01 6.03798211e-01
6.55175269e-01 1.98881910e-03 5.15625417e-01 1.12408483e+00
5.91316104e-01 2.48295084e-01 1.75924331e-01 7.30041623e-01
1.94264099e-01 -2.82489389e-01 -6.82080567e-01 2.91365087e-01
-1.56979412e-02 1.15289426e+00 -3.08296889e-01 -5.79670556e-02
-3.02623749e-01 1.13356948e+00 3.54686499e-01 1.44790232e-01
-8.54152620e-01 -3.72664817e-02 5.77200651e-01 1.17616408e-01
2.63613731e-01 -5.14979005e-01 -6.30543947e-01 -1.12889099e+00
2.18292415e-01 -6.02186024e-01 6.25527203e-01 -1.02587378e+00
-1.30722713e+00 7.55116582e-01 8.26707929e-02 -1.38174248e+00
5.01879342e-02 -4.36264426e-01 -4.34667081e-01 1.15536129e+00
-1.99609923e+00 -1.82925546e+00 -4.36677456e-01 6.85003400e-01
4.95432824e-01 -1.76881403e-02 7.01455176e-01 4.62243140e-01
-4.48957324e-01 1.84524432e-01 -1.27861788e-02 1.73929542e-01
3.67337018e-01 -1.12108028e+00 4.09794837e-01 1.30338085e+00
-9.82381552e-02 1.77269697e-01 2.73905635e-01 -5.23600042e-01
-1.06372881e+00 -1.46183932e+00 1.13666010e+00 -2.45019402e-02
5.48364222e-01 -8.37768391e-02 -9.23709989e-01 3.83002967e-01
9.93066207e-02 -2.91801449e-02 5.21511912e-01 -3.47517818e-01
-4.37871814e-01 -4.63250875e-01 -1.08755016e+00 2.48255804e-01
1.02891850e+00 -8.78581226e-01 -6.17547929e-01 1.67133674e-01
3.35164100e-01 -2.68614620e-01 -1.14123452e+00 4.46389109e-01
3.98988634e-01 -1.09500849e+00 1.22143710e+00 -2.18327530e-02
8.11988294e-01 -4.94495511e-01 -5.88096380e-01 -1.34214962e+00
-4.60601479e-01 1.15553327e-02 3.47579747e-01 9.65095282e-01
2.11872116e-01 -4.17526931e-01 3.02851528e-01 1.83723256e-01
-6.88991845e-01 -5.76458573e-01 -1.26617289e+00 -4.66015995e-01
2.46917188e-01 -7.09722698e-01 6.64140224e-01 5.57292640e-01
-3.24241877e-01 2.18530267e-01 -6.10966802e-01 6.06757104e-01
7.64663994e-01 1.61839560e-01 2.60854326e-02 -1.17891431e+00
2.73515657e-02 -4.14305776e-01 -3.51723552e-01 -9.90588665e-01
1.31484792e-01 -1.00657094e+00 1.12636320e-01 -1.85784340e+00
1.62447751e-01 -3.08726728e-01 -3.46687794e-01 9.18011725e-01
-1.43067375e-01 4.93500412e-01 4.09169570e-02 2.00340569e-01
-7.17198104e-02 9.28773046e-01 1.11926150e+00 -6.01960123e-01
2.43137583e-01 -2.56613284e-01 -5.11436343e-01 4.58482862e-01
9.47760820e-01 -4.33168054e-01 -3.65569830e-01 -1.26704168e+00
2.40743697e-01 1.16950490e-01 1.02611077e+00 -1.25251436e+00
2.65806645e-01 -3.61208357e-02 1.93731636e-01 -9.03234422e-01
2.37107873e-01 -9.66936350e-01 1.48524791e-01 5.29044509e-01
-1.64569810e-01 6.54040650e-02 7.27187470e-03 4.55860347e-01
-6.70200825e-01 1.70392692e-01 6.58026874e-01 -1.63567230e-01
-9.33023095e-01 6.13767087e-01 -6.06261969e-01 -4.09273893e-01
7.43876994e-01 -2.07537338e-01 -2.83064425e-01 -1.64942011e-01
-6.39858961e-01 1.33891508e-01 1.07319228e-01 2.93591321e-01
1.04131448e+00 -1.24417996e+00 -1.22781575e+00 1.07108921e-01
1.99157640e-01 3.07694674e-01 5.97653270e-01 9.84841466e-01
-6.35331869e-01 3.79226387e-01 -3.64504009e-01 -5.40046513e-01
-8.88499677e-01 1.98373154e-01 5.39047718e-01 -6.02174282e-01
-7.08348155e-01 7.76562572e-01 1.53434530e-01 -3.46197277e-01
-2.92846680e-01 -5.81176281e-01 -2.47501433e-01 3.60148728e-01
5.26563525e-01 3.69427383e-01 4.79439765e-01 -1.05026650e+00
-5.25550663e-01 6.24575794e-01 2.60144174e-01 -3.82924795e-01
1.84971130e+00 -3.95142078e-01 -4.74665225e-01 2.01517597e-01
1.11933982e+00 -6.00441694e-01 -1.25063765e+00 -4.23952192e-01
-9.94213894e-02 -1.84382647e-01 5.65984368e-01 -8.66258800e-01
-1.39934719e+00 1.48176551e+00 7.43701160e-01 -1.10960074e-01
1.78235304e+00 -1.85005799e-01 9.55941141e-01 5.07687271e-01
5.42098165e-01 -9.29737926e-01 -1.52039593e-02 7.91437209e-01
1.35439003e+00 -1.17374158e+00 -1.03542730e-01 -4.85525690e-02
-5.46864331e-01 1.08596623e+00 3.54703873e-01 1.06571525e-01
8.59297216e-01 1.76024109e-01 -3.58197279e-02 -5.66670716e-01
-4.15717155e-01 -5.53825498e-01 3.38265300e-01 6.38151526e-01
4.02647436e-01 1.75063461e-01 -1.58269361e-01 4.35304195e-01
-2.40950263e-03 -2.77092587e-02 2.26975486e-01 9.31675851e-01
-7.09350407e-01 -9.16745484e-01 -4.43315804e-01 3.40503633e-01
-2.77283818e-01 -4.86631304e-01 -2.23834351e-01 2.14755684e-01
2.48189762e-01 9.48155463e-01 -9.69946459e-02 -3.39684695e-01
4.16849196e-01 -2.46772274e-01 1.97710678e-01 -3.32360297e-01
-9.23142791e-01 -1.61882296e-01 -3.04713473e-02 -5.70687234e-01
-7.01223493e-01 -7.07672894e-01 -1.11992788e+00 -5.60580529e-02
2.38685638e-01 -3.17586601e-01 3.45139682e-01 1.05485380e+00
1.24347694e-01 5.86230636e-01 6.24793231e-01 -1.04016006e+00
-5.44018865e-01 -1.04969597e+00 -6.74766421e-01 1.60089046e-01
5.39992809e-01 -5.67781687e-01 -2.08566468e-02 1.83797535e-02]
|
[9.69180679321289, -1.2826253175735474]
|
426a2276-9025-4825-ae05-97cf5b7dfbea
|
searching-for-a-search-method-benchmarking
|
2009.06368
| null |
https://arxiv.org/abs/2009.06368v2
|
https://arxiv.org/pdf/2009.06368v2.pdf
|
Searching for a Search Method: Benchmarking Search Algorithms for Generating NLP Adversarial Examples
|
We study the behavior of several black-box search algorithms used for generating adversarial examples for natural language processing (NLP) tasks. We perform a fine-grained analysis of three elements relevant to search: search algorithm, search space, and search budget. When new search algorithms are proposed in past work, the attack search space is often modified alongside the search algorithm. Without ablation studies benchmarking the search algorithm change with the search space held constant, one cannot tell if an increase in attack success rate is a result of an improved search algorithm or a less restrictive search space. Additionally, many previous studies fail to properly consider the search algorithms' run-time cost, which is essential for downstream tasks like adversarial training. Our experiments provide a reproducible benchmark of search algorithms across a variety of search spaces and query budgets to guide future research in adversarial NLP. Based on our experiments, we recommend greedy attacks with word importance ranking when under a time constraint or attacking long inputs, and either beam search or particle swarm optimization otherwise. Code implementation shared via https://github.com/QData/TextAttack-Search-Benchmark
|
['Yanjun Qi', 'John X. Morris', 'Eli Lifland', 'Jin Yong Yoo']
|
2020-09-09
| null |
https://aclanthology.org/2020.blackboxnlp-1.30
|
https://aclanthology.org/2020.blackboxnlp-1.30.pdf
|
emnlp-blackboxnlp-2020-11
|
['adversarial-text']
|
['adversarial']
|
[ 4.81626302e-01 -1.32483646e-01 -2.51675040e-01 -8.29427838e-02
-1.09065676e+00 -1.29180741e+00 8.79247010e-01 1.48323685e-01
-6.52060926e-01 7.03007400e-01 4.30653960e-01 -8.38945746e-01
-2.10813418e-01 -8.28352749e-01 -4.59164143e-01 -5.87732673e-01
2.32949257e-01 7.53334939e-01 2.01641992e-01 -4.02407050e-01
3.87871742e-01 7.34716415e-01 -9.69910860e-01 4.89443466e-02
5.84928989e-01 5.20129442e-01 -2.18145728e-01 1.02870262e+00
1.24322236e-01 4.72948313e-01 -1.17510903e+00 -7.52510726e-01
7.03763902e-01 -4.45638806e-01 -8.74510944e-01 -9.60684478e-01
3.80212575e-01 -5.22335172e-02 -4.95809227e-01 1.29813135e+00
1.07693434e+00 3.10416758e-01 5.25509536e-01 -1.44475746e+00
-4.87782001e-01 9.11377072e-01 -3.27628828e-03 8.04098785e-01
3.52329940e-01 6.66847467e-01 1.19368267e+00 -4.02754396e-01
6.75597906e-01 1.26856828e+00 5.25381505e-01 8.35320413e-01
-1.23826993e+00 -1.01985621e+00 -1.85065251e-03 8.74474123e-02
-1.10800993e+00 -4.81239289e-01 5.75831652e-01 -1.95322379e-01
1.02440786e+00 8.61911833e-01 2.64026850e-01 1.55741954e+00
2.40620285e-01 3.98403347e-01 7.25636780e-01 -2.82857180e-01
3.74852836e-01 1.78074148e-02 -1.69064656e-01 5.17723739e-01
3.28042626e-01 7.56239653e-01 -4.13048387e-01 -7.98548818e-01
3.14344734e-01 -3.50334138e-01 -3.50221246e-01 2.45057922e-02
-9.66546476e-01 1.24730182e+00 3.54122877e-01 3.92658502e-01
-1.44744381e-01 4.47851896e-01 4.96724576e-01 5.95737696e-01
3.45189810e-01 1.43127024e+00 -5.95290005e-01 -3.22326094e-01
-9.57251370e-01 6.75418973e-01 1.07136238e+00 4.99677241e-01
1.97922260e-01 7.77945071e-02 -4.91614550e-01 4.48859870e-01
-4.68847789e-02 5.09049773e-01 6.58342183e-01 -8.95264745e-01
6.30892396e-01 1.47418365e-01 8.84215608e-02 -9.79312897e-01
-3.07371795e-01 -4.37568128e-01 -4.51951414e-01 3.07531834e-01
7.00613022e-01 -4.10975367e-01 -9.56042707e-01 1.74752414e+00
1.08987138e-01 -1.17605276e-01 9.07792971e-02 8.45555305e-01
5.52950859e-01 5.08926213e-01 2.02726170e-01 -1.31083027e-01
1.33895862e+00 -8.87215197e-01 -3.74268234e-01 -5.47747195e-01
8.92339110e-01 -1.01320565e+00 1.18730843e+00 3.07559788e-01
-1.13788104e+00 -5.59381731e-02 -8.89022768e-01 2.60214478e-01
-6.33702934e-01 -6.10527217e-01 5.27556002e-01 9.89703834e-01
-7.74632394e-01 8.32067609e-01 -5.81479728e-01 -9.71071869e-02
3.09114695e-01 2.11997494e-01 1.19210511e-01 1.67572483e-01
-1.76912200e+00 1.07029033e+00 1.40859097e-01 -3.49697173e-01
-8.31198812e-01 -9.31915879e-01 -5.38995504e-01 9.14622396e-02
3.03875804e-01 -8.64891291e-01 1.31399155e+00 -5.69783807e-01
-1.24179351e+00 5.09930670e-01 8.47962126e-02 -8.03685129e-01
7.73079872e-01 -1.97107151e-01 -3.61314833e-01 -4.70615253e-02
-2.26891711e-02 2.86466330e-01 9.10896361e-01 -9.48116660e-01
-2.93993860e-01 -1.04121529e-01 2.79161364e-01 3.79421562e-01
-1.86860412e-01 3.81389469e-01 -8.97551775e-02 -1.16440845e+00
-5.08733273e-01 -1.07875681e+00 -5.57546139e-01 -2.63007849e-01
-4.48558390e-01 4.94959280e-02 5.33494949e-01 -4.04290199e-01
1.52685344e+00 -2.09727120e+00 1.85146451e-01 4.95142132e-01
1.12370320e-01 4.63527471e-01 -3.26752663e-01 6.16410673e-01
-2.36659691e-01 8.30680311e-01 -6.88118786e-02 1.06048219e-01
2.92067468e-01 -1.79618880e-01 -7.85583079e-01 3.16807061e-01
-1.28785357e-01 9.61503327e-01 -9.91815329e-01 -4.17834133e-01
-1.02512263e-01 4.66420203e-02 -7.39859462e-01 4.42839079e-02
-3.22605401e-01 2.04771265e-01 -7.34756470e-01 5.28434575e-01
2.10049406e-01 -1.07381731e-01 -2.35731810e-01 -5.83334407e-03
3.04196298e-01 5.03045678e-01 -1.01078892e+00 1.36316669e+00
-5.29108703e-01 4.87354696e-01 3.99854742e-02 -5.55498242e-01
5.28307080e-01 2.22942337e-01 1.16312653e-01 -6.56832218e-01
-6.07343428e-02 2.25805670e-01 2.97270089e-01 -2.98741192e-01
3.80282402e-01 5.92488647e-02 -2.82314450e-01 7.96894789e-01
-3.75620723e-01 -6.11621976e-01 2.48959646e-01 1.87127620e-01
1.79970169e+00 -4.11893606e-01 1.12708747e-01 -6.11145161e-02
2.46456623e-01 3.18108529e-01 4.17465478e-01 1.20515084e+00
-3.57495397e-01 4.94322419e-01 4.24716055e-01 -4.60581183e-01
-8.85080218e-01 -1.01515591e+00 -8.97150263e-02 1.47311199e+00
6.05443604e-02 -6.01468980e-01 -6.27451539e-01 -8.87909472e-01
8.44850242e-02 9.41843510e-01 -6.01615787e-01 -5.03081679e-01
-8.19072664e-01 -9.64477003e-01 1.24199462e+00 1.83510780e-01
2.11141676e-01 -1.38389289e+00 -6.36436939e-01 -3.90847102e-02
4.27708483e-06 -6.19606555e-01 -9.10872936e-01 2.73929060e-01
-3.96965265e-01 -9.48650777e-01 -4.31803554e-01 -4.74484771e-01
5.30579150e-01 -2.82984853e-01 1.23493421e+00 3.10284168e-01
-4.64864105e-01 3.58942151e-01 -4.33564633e-01 -5.70422173e-01
-8.77161980e-01 2.57009476e-01 6.97657559e-03 -7.88658261e-01
-1.44719090e-02 -4.06802952e-01 -7.70302594e-01 3.20889503e-01
-9.03952181e-01 -5.97147882e-01 5.10699511e-01 1.08240485e+00
2.37463325e-01 3.05251211e-01 3.36934417e-01 -1.07203698e+00
1.32695580e+00 -5.02072155e-01 -6.59398913e-01 4.74688143e-01
-8.05374086e-01 4.56461012e-01 6.49599552e-01 -7.00858831e-01
-5.71273327e-01 -1.62030667e-01 -2.28748575e-01 -5.38051128e-01
-2.44119577e-03 3.05338621e-01 -4.91245762e-02 -7.75524154e-02
1.33130991e+00 8.76396745e-02 -2.63631731e-01 -6.09453134e-02
3.20724696e-01 1.45702541e-01 2.04045832e-01 -7.81690717e-01
1.36936390e+00 3.80723439e-02 -2.03518569e-01 -1.77636474e-01
-5.40939212e-01 -1.19211853e-01 2.32879966e-02 2.81496346e-01
7.81293392e-01 -2.72101402e-01 -5.66494644e-01 9.28010270e-02
-1.13380861e+00 -5.30330956e-01 -5.07367969e-01 1.96027786e-01
-3.53848755e-01 2.57481903e-01 -4.92554724e-01 -5.75320303e-01
-8.77871454e-01 -1.43480277e+00 8.66228104e-01 3.76571044e-02
-7.67196178e-01 -1.14253223e+00 2.59402275e-01 4.20990050e-01
6.19995296e-01 1.45970181e-01 9.24044013e-01 -1.28101671e+00
-3.10333848e-01 -4.02746618e-01 2.23577097e-01 -5.81614040e-02
-4.46945876e-02 -1.96264297e-01 -8.41785789e-01 -4.55720514e-01
1.49306571e-02 -3.06974471e-01 6.67289972e-01 2.16091365e-01
1.15009546e+00 -8.20506692e-01 -4.10740137e-01 8.87232602e-01
1.05827296e+00 3.45512420e-01 3.70796710e-01 6.35860503e-01
3.19618523e-01 4.23228532e-01 7.08934546e-01 2.84284055e-01
-4.18234199e-01 6.37898088e-01 4.71486121e-01 3.09209656e-02
1.96028888e-01 -2.99252659e-01 3.61589491e-01 -1.88142769e-02
3.40929836e-01 -7.69550025e-01 -1.24413323e+00 2.47162923e-01
-1.47473705e+00 -1.33298218e+00 6.55574501e-01 2.18763876e+00
1.20794868e+00 5.94803214e-01 7.41059110e-02 -2.23009977e-02
5.88821769e-01 4.45436776e-01 -7.90101945e-01 -7.50352144e-01
3.86753343e-02 4.79377985e-01 8.18844378e-01 7.69863129e-01
-9.30925727e-01 1.08847129e+00 6.90316916e+00 1.23785722e+00
-1.09562457e+00 -8.42457265e-02 5.78807712e-01 -4.64308232e-01
-5.38884282e-01 -7.13958293e-02 -6.93624616e-01 6.34121299e-01
1.06876826e+00 -8.47205639e-01 8.23755860e-01 8.71203005e-01
-6.87906817e-02 3.68948013e-01 -1.13648248e+00 7.08125651e-01
-1.80519596e-01 -1.49721408e+00 8.57745409e-02 9.82874259e-02
3.85419697e-01 1.41507849e-01 4.46524739e-01 2.41595119e-01
9.82589543e-01 -1.37902558e+00 5.33372521e-01 -4.48881201e-02
6.08068049e-01 -6.36984050e-01 5.63867748e-01 2.95857280e-01
-7.09620118e-01 -2.79581636e-01 6.28888085e-02 2.92987734e-01
6.56236187e-02 2.17456311e-01 -1.12310469e+00 1.28169596e-01
5.88859916e-01 -4.39799912e-02 -5.54301798e-01 7.13498473e-01
-2.56920427e-01 7.90095925e-01 -4.97606128e-01 -3.57921183e-01
2.51045883e-01 3.75251472e-01 1.16576326e+00 1.22081649e+00
8.40908885e-02 8.59248489e-02 5.13814948e-02 8.50008667e-01
-4.42746989e-02 -9.30328015e-03 -6.23591542e-01 -4.04661626e-01
1.16097558e+00 9.77641225e-01 -5.39906800e-01 -2.30625365e-02
2.31131375e-01 8.19315970e-01 -9.15847570e-02 4.23142105e-01
-1.05959380e+00 -4.71581995e-01 9.32914376e-01 5.54907843e-02
1.07733291e-02 4.74827886e-02 -3.54349047e-01 -7.65062332e-01
-7.02127144e-02 -1.45611560e+00 7.13543653e-01 -4.50634807e-01
-1.29927886e+00 8.39900255e-01 1.40329838e-01 -9.86885786e-01
-5.86723447e-01 -3.47994298e-01 -8.96213174e-01 1.03524554e+00
-9.71566021e-01 -6.06292725e-01 2.76328057e-01 5.04181802e-01
6.52741134e-01 -5.82559407e-01 1.00106573e+00 -2.88177375e-02
-4.60813344e-01 1.22104979e+00 -7.09262043e-02 1.65073723e-01
8.29992115e-01 -1.18048096e+00 9.24051166e-01 9.50484753e-01
3.36946815e-01 8.47926676e-01 1.26120698e+00 -6.84986591e-01
-1.22719765e+00 -8.56567919e-01 6.18086457e-01 -9.32505548e-01
9.96203005e-01 -2.85030514e-01 -5.99237025e-01 2.08254382e-01
4.66210172e-02 -1.05703384e-01 6.18496180e-01 1.60419986e-01
-5.43479085e-01 1.96536466e-01 -1.38320422e+00 9.71721590e-01
1.07535589e+00 -4.42359954e-01 -4.05372441e-01 6.60698116e-01
1.07563305e+00 -8.07194054e-01 -7.78677464e-01 3.59838575e-01
3.02983016e-01 -4.39106137e-01 1.23337567e+00 -8.49905968e-01
2.33550787e-01 8.53611529e-02 -7.40680918e-02 -1.54468393e+00
-2.87490457e-01 -1.04654038e+00 -7.61293918e-02 7.65032411e-01
8.80663395e-01 -7.95631289e-01 9.44101810e-01 8.56392980e-01
1.10497907e-01 -9.21672225e-01 -9.42862451e-01 -7.60899544e-01
4.49822217e-01 -3.79943579e-01 7.60338545e-01 7.94207931e-01
-1.40222624e-01 3.18982273e-01 -8.37987438e-02 1.93153143e-01
3.35507572e-01 -3.99567150e-02 6.36675537e-01 -7.69910991e-01
-5.74881494e-01 -9.64695334e-01 -2.00163379e-01 -5.08608162e-01
2.82699704e-01 -8.69876742e-01 -1.43216848e-01 -1.17231500e+00
-2.79125273e-01 -4.73671049e-01 -1.89812794e-01 6.66112542e-01
-3.91015381e-01 -2.60959771e-02 2.97952563e-01 3.08485359e-01
-2.23112777e-01 1.31974936e-01 9.66596246e-01 -3.94490242e-01
-3.23707998e-01 3.02414924e-01 -7.90909350e-01 4.69545156e-01
1.14829361e+00 -8.03596795e-01 -7.00932145e-01 -5.98463535e-01
5.39698541e-01 -8.41811448e-02 3.22877944e-01 -7.58026958e-01
4.89447564e-01 -4.39547807e-01 1.10135093e-01 2.10300945e-02
2.87016630e-01 -7.15573549e-01 6.18155412e-02 9.02207732e-01
-9.38947916e-01 4.72181827e-01 4.28425610e-01 5.72335184e-01
1.76253110e-01 -4.84964997e-01 7.15590179e-01 -5.43452390e-02
-3.26859206e-01 2.71845162e-01 -2.31718853e-01 6.22006893e-01
9.45583642e-01 -3.53294052e-02 -6.28633797e-01 -4.06863034e-01
-7.63378501e-01 2.32622281e-01 4.44216698e-01 6.48398042e-01
4.58591819e-01 -1.04184675e+00 -8.69109452e-01 -4.51248139e-03
-1.26246333e-01 -3.38682443e-01 -1.91270500e-01 1.87899962e-01
-6.45748973e-01 3.06940198e-01 1.62943482e-01 1.08808242e-01
-1.37716770e+00 5.97453833e-01 4.81398344e-01 -7.17185020e-01
-3.35898668e-01 1.27530742e+00 -1.32176414e-01 -3.69715363e-01
4.95503247e-01 9.23250839e-02 1.85990795e-01 -1.04085416e-01
4.07581210e-01 3.08029592e-01 -4.06494811e-02 -1.31549731e-01
-4.12769973e-01 1.44787595e-01 -1.95798948e-01 -5.14918625e-01
9.70132649e-01 4.23051685e-01 2.09687918e-01 -9.25302580e-02
9.35145795e-01 3.31389368e-01 -5.89361608e-01 -1.56894773e-01
-1.14568494e-01 -6.08125091e-01 -8.07759166e-02 -1.01888859e+00
-9.58657146e-01 3.61118585e-01 4.63817358e-01 4.04250562e-01
9.82624471e-01 -1.13051593e-01 7.66454279e-01 6.27239227e-01
2.56902844e-01 -9.28605437e-01 8.74877498e-02 3.96058917e-01
1.26059461e+00 -9.99940813e-01 2.59356022e-01 -1.03818059e-01
-6.42034650e-01 6.83257639e-01 5.80579698e-01 -3.57209407e-02
6.91356361e-01 5.71025610e-01 7.09842518e-02 -1.40533179e-01
-9.24035907e-01 1.39370754e-01 1.74931780e-01 3.84046942e-01
1.10826500e-01 -1.71298727e-01 -2.93782681e-01 4.47816342e-01
-8.83923948e-01 -6.22904181e-01 2.66944524e-02 8.07171524e-01
-2.05913231e-01 -1.39531493e+00 -4.41337556e-01 5.30186415e-01
-8.20553362e-01 -6.29570663e-01 -7.24255323e-01 7.40042686e-01
-1.60753563e-01 1.02287245e+00 -1.73927635e-01 -5.46832502e-01
3.78636450e-01 4.10041474e-02 2.36971691e-01 -6.13611102e-01
-1.09935498e+00 -5.40246785e-01 2.78253675e-01 -6.78552806e-01
2.29211122e-01 -6.96779609e-01 -1.09233117e+00 -4.58047658e-01
-4.33896184e-01 5.62638819e-01 5.85222602e-01 6.39935493e-01
3.62865955e-01 2.92442590e-01 4.70385253e-01 -4.04314786e-01
-1.10819924e+00 -6.39718056e-01 1.26530245e-01 4.45405543e-01
1.78023398e-01 -2.81196058e-01 -8.55465949e-01 -1.62681490e-01]
|
[6.034277439117432, 8.134843826293945]
|
6d7b4ff4-7968-4656-9ba4-15c545b710d0
|
hierarchical-task-network-planning-for
|
2306.08359
| null |
https://arxiv.org/abs/2306.08359v1
|
https://arxiv.org/pdf/2306.08359v1.pdf
|
Hierarchical Task Network Planning for Facilitating Cooperative Multi-Agent Reinforcement Learning
|
Exploring sparse reward multi-agent reinforcement learning (MARL) environments with traps in a collaborative manner is a complex task. Agents typically fail to reach the goal state and fall into traps, which affects the overall performance of the system. To overcome this issue, we present SOMARL, a framework that uses prior knowledge to reduce the exploration space and assist learning. In SOMARL, agents are treated as part of the MARL environment, and symbolic knowledge is embedded using a tree structure to build a knowledge hierarchy. The framework has a two-layer hierarchical structure, comprising a hybrid module with a Hierarchical Task Network (HTN) planning and meta-controller at the higher level, and a MARL-based interactive module at the lower level. The HTN module and meta-controller use Hierarchical Domain Definition Language (HDDL) and the option framework to formalize symbolic knowledge and obtain domain knowledge and a symbolic option set, respectively. Moreover, the HTN module leverages domain knowledge to guide low-level agent exploration by assisting the meta-controller in selecting symbolic options. The meta-controller further computes intrinsic rewards of symbolic options to limit exploration behavior and adjust HTN planning solutions as needed. We evaluate SOMARL on two benchmarks, FindTreasure and MoveBox, and report superior performance over state-of-the-art MARL and subgoal-based baselines for MARL environments significantly.
|
['Jianye Hao', 'Chao Yu', 'Kai Zhang', 'Chen Chen', 'Hankz Hankui Zhuo', 'Xuechen Mu']
|
2023-06-14
| null | null | null | null |
['multi-agent-reinforcement-learning']
|
['methodology']
|
[-3.09132665e-01 1.96049497e-01 -4.98316407e-01 2.89446209e-02
-8.29383075e-01 -6.94573104e-01 5.01846433e-01 3.06672275e-01
-6.73308372e-01 9.84835386e-01 1.72348782e-01 -3.01141977e-01
-3.20812374e-01 -1.04170740e+00 -5.86945951e-01 -6.51528299e-01
-5.92305481e-01 9.29794848e-01 5.76647937e-01 -5.54728866e-01
2.77258843e-01 1.23555854e-01 -1.38657534e+00 1.88476562e-01
1.11536801e+00 7.76603580e-01 1.77419707e-01 3.57487768e-01
3.24422829e-02 1.31656575e+00 -5.48521340e-01 3.50446701e-01
4.30240184e-01 -2.10947350e-01 -1.02963841e+00 -3.82396281e-01
-3.28562379e-01 -4.46352065e-01 -1.42648354e-01 8.66905928e-01
4.34162676e-01 6.86555982e-01 2.19251215e-01 -1.48209333e+00
-2.22547308e-01 1.24318182e+00 -4.19375777e-01 -1.90363638e-02
3.94910127e-01 9.12614524e-01 9.53419983e-01 -5.09520352e-01
7.21196175e-01 1.67895448e+00 2.95574605e-01 4.79679674e-01
-1.28707838e+00 -6.32714748e-01 7.16667235e-01 1.68747142e-01
-1.06321192e+00 -2.36160353e-01 5.72393060e-01 -1.63612485e-01
1.50731075e+00 -1.30045652e-01 8.86978626e-01 9.24519002e-01
2.48390257e-01 7.67460942e-01 1.12143588e+00 -8.33896548e-02
9.30026412e-01 -3.05935264e-01 -6.54206276e-02 9.41951036e-01
5.22291958e-02 7.24793911e-01 -7.52161980e-01 -3.86191308e-01
1.02982497e+00 -3.19601297e-01 3.14659953e-01 -7.80550718e-01
-1.08926928e+00 1.22591162e+00 8.29195261e-01 -1.66302785e-01
-6.64387703e-01 3.95325541e-01 5.05502045e-01 4.36258405e-01
-3.39698076e-01 9.63327050e-01 -3.06966305e-01 -4.53852057e-01
-4.67624307e-01 7.14607298e-01 9.57367182e-01 8.07343721e-01
1.03745842e+00 8.59159380e-02 -2.73006350e-01 4.74618673e-01
4.14801568e-01 3.34926516e-01 3.81861418e-01 -1.54925323e+00
5.25439918e-01 9.22043145e-01 3.63254547e-01 -4.83921230e-01
-6.48825049e-01 -4.28914636e-01 -2.34299656e-02 9.60169792e-01
1.42927662e-01 -4.87289548e-01 -1.05018938e+00 1.85612023e+00
6.27345085e-01 6.74643926e-03 5.37804186e-01 1.05313742e+00
6.69561565e-01 4.90938663e-01 3.14394176e-01 -8.47044773e-03
1.04694128e+00 -1.51639318e+00 -2.50267565e-01 -6.27233267e-01
9.46104228e-01 1.12394780e-01 1.08985388e+00 4.87582475e-01
-1.26035392e+00 -1.32637516e-01 -9.67876673e-01 1.09983601e-01
-3.67465615e-01 -2.93820530e-01 7.96781182e-01 1.74688861e-01
-1.15325582e+00 5.09583950e-01 -1.11681259e+00 -1.18779158e-02
4.82104361e-01 5.50027132e-01 3.87520753e-02 1.35180587e-02
-1.30672228e+00 1.09857631e+00 8.74377668e-01 -1.43949732e-01
-1.54303122e+00 -3.18697929e-01 -1.16537046e+00 5.60695790e-02
1.04987240e+00 -6.31338775e-01 1.59074461e+00 -4.26381171e-01
-1.92396748e+00 2.51811370e-02 4.22909975e-01 -6.78515971e-01
4.36786741e-01 -3.64989825e-02 5.39361574e-02 2.87842844e-02
3.32031012e-01 1.02383637e+00 6.08901739e-01 -1.20145500e+00
-9.83874083e-01 1.49682853e-02 5.89176416e-01 7.98788846e-01
3.12157243e-01 -4.95645732e-01 -3.60573620e-01 -9.08289552e-02
-1.99939758e-01 -9.62449968e-01 -8.79743993e-01 -5.59018373e-01
-2.19013810e-01 -4.86361682e-01 6.00012362e-01 -1.13009103e-01
1.19689453e+00 -1.84379959e+00 4.50764954e-01 5.60851097e-01
2.71233857e-01 -8.88659582e-02 -7.40863621e-01 5.47059000e-01
4.99515384e-01 -1.20839924e-01 7.20209032e-02 -2.73411095e-01
3.26475739e-01 5.92896521e-01 -2.32033804e-01 6.14879839e-02
8.06551352e-02 1.04838860e+00 -1.44139934e+00 -5.23429811e-01
2.61483788e-01 -7.40998015e-02 -9.55903411e-01 3.55113149e-02
-1.02857625e+00 3.77553076e-01 -9.32834148e-01 9.25385594e-01
2.79290199e-01 -1.71474099e-01 4.57781971e-01 4.15200293e-01
-5.07141590e-01 6.89086020e-01 -1.38811195e+00 1.90831828e+00
-3.17112803e-01 -3.28992084e-02 2.00626269e-01 -4.22342181e-01
7.95357347e-01 -1.41851112e-01 3.74931872e-01 -9.16415811e-01
2.95942090e-02 1.48967549e-01 1.54962584e-01 -1.87314183e-01
8.07504296e-01 4.85528946e-01 -4.10719365e-01 4.32427913e-01
-2.70537674e-01 -6.34911060e-02 5.60475826e-01 1.85406700e-01
1.52847350e+00 6.79016888e-01 3.08405280e-01 -3.47636268e-02
2.76392847e-01 7.32278705e-01 8.64049733e-01 1.28051066e+00
-1.77820489e-01 -5.89974225e-01 8.03939164e-01 -5.58346272e-01
-5.44893861e-01 -1.07600713e+00 5.09545863e-01 1.59668922e+00
4.61173147e-01 -5.40149331e-01 -4.72663552e-01 -8.61431837e-01
3.21629763e-01 8.80655408e-01 -7.56574273e-01 -1.48246795e-01
-8.61383021e-01 -2.23941028e-01 4.70951706e-01 5.66997111e-01
6.88618124e-01 -1.75440323e+00 -1.50462759e+00 5.54614067e-01
6.03624173e-02 -7.44073212e-01 -2.22847342e-01 7.06883311e-01
-6.43672526e-01 -1.01817632e+00 7.93292373e-02 -5.81262171e-01
4.25276726e-01 -2.48539135e-01 1.24735260e+00 -1.48955814e-03
-5.70418835e-02 2.46040076e-01 -4.71429169e-01 -2.98802201e-02
-4.30106699e-01 3.98856372e-01 -3.81866768e-02 -8.53223443e-01
2.15205863e-01 -5.60516596e-01 -4.84005272e-01 3.22831780e-01
-5.04919469e-01 4.99409698e-02 7.74765074e-01 8.63761008e-01
5.05107284e-01 1.26860201e-01 5.68730354e-01 -4.70684946e-01
8.59533846e-01 -5.63569009e-01 -1.24433494e+00 -5.98944947e-02
-5.67622006e-01 4.14774537e-01 5.09416640e-01 -3.80252868e-01
-7.78237283e-01 1.92101985e-01 3.44179720e-01 -3.87297571e-01
2.70412099e-02 9.26309764e-01 -2.00727563e-02 -9.90057588e-02
8.58086228e-01 1.06457584e-01 1.69906512e-01 -1.63060769e-01
5.35368741e-01 9.44329798e-02 4.86860782e-01 -1.04034555e+00
7.61068761e-01 8.70992914e-02 1.00873532e-02 -1.30061567e-01
-5.54997206e-01 2.13797856e-02 -1.58716202e-01 -1.68663844e-01
5.00143111e-01 -9.30321634e-01 -1.32427645e+00 1.05896391e-01
-7.56607711e-01 -1.44600189e+00 -7.07893848e-01 2.24965677e-01
-8.99472475e-01 -2.22702071e-01 -5.10589898e-01 -7.30497777e-01
-7.42240325e-02 -1.66778159e+00 6.90093994e-01 4.76532966e-01
-3.46614122e-02 -8.04382563e-01 3.41433793e-01 6.11284235e-03
4.06262398e-01 3.70171905e-01 9.39843774e-01 -6.80455565e-01
-1.01422596e+00 6.12477183e-01 4.78981435e-02 -4.82200146e-01
-2.36810073e-01 -6.44542515e-01 -4.45381016e-01 -5.53986013e-01
-5.95527709e-01 -9.17699158e-01 8.55278432e-01 3.87231827e-01
5.66590607e-01 -3.22732359e-01 -5.82397580e-01 4.48394179e-01
1.25253773e+00 4.11082715e-01 4.54011858e-01 1.12726200e+00
2.38502905e-01 3.51002425e-01 1.15605187e+00 7.65149951e-01
8.50653052e-01 6.73445880e-01 9.99969602e-01 2.12707594e-01
2.24449128e-01 -3.58913600e-01 7.35610008e-01 -2.40837201e-03
1.15063496e-01 -9.92943496e-02 -1.06887412e+00 3.67819130e-01
-2.46783328e+00 -8.95214677e-01 4.98999298e-01 2.00965571e+00
1.19276845e+00 1.96389154e-01 5.82897544e-01 -4.10931677e-01
1.89737260e-01 2.56684661e-01 -1.29130709e+00 -4.58469093e-01
2.99561560e-01 -1.96000725e-01 5.41743994e-01 9.63568389e-01
-1.02422297e+00 1.75294113e+00 5.77444792e+00 6.95936024e-01
-8.30190003e-01 -1.33761689e-01 -4.25723903e-02 -3.39609742e-01
-7.25726709e-02 1.63175270e-01 -1.01036060e+00 4.89633307e-02
6.79304004e-01 -1.94258783e-02 1.16745043e+00 1.07146096e+00
1.02785490e-01 -4.49510843e-01 -1.23695099e+00 4.90548760e-01
-6.06043398e-01 -1.38189518e+00 -1.98376879e-01 2.73651272e-01
5.92158437e-01 4.58481371e-01 2.54849225e-01 1.08534324e+00
1.50670409e+00 -1.27429605e+00 7.04609871e-01 2.52338916e-01
3.60507399e-01 -1.13153625e+00 3.45072865e-01 4.72970903e-01
-1.40411294e+00 -6.34920239e-01 -2.52523303e-01 -3.33920747e-01
4.90928814e-02 -2.15860426e-01 -1.00970018e+00 4.56403375e-01
6.62703633e-01 6.35271549e-01 -4.28474009e-01 9.91294503e-01
-5.56761920e-01 8.16718712e-02 -4.11087364e-01 -1.88197955e-01
9.87385392e-01 -1.30483747e-01 8.40344489e-01 7.25419164e-01
-2.73770601e-01 5.20984940e-02 1.06335592e+00 1.14129972e+00
1.87383145e-01 -4.73333150e-01 -3.13454509e-01 -6.03764430e-02
9.87175584e-01 1.20049810e+00 -5.56168735e-01 -1.76353052e-01
-9.54138413e-02 5.12136161e-01 8.36263776e-01 5.96270561e-01
-8.83349121e-01 -2.31955320e-01 8.50939810e-01 -3.25151920e-01
3.08719188e-01 -4.05938447e-01 -1.82068929e-01 -6.87588871e-01
-3.87482911e-01 -1.31748974e+00 5.76377928e-01 -4.79925871e-01
-8.28054249e-01 6.66455328e-01 3.23047601e-02 -8.74169111e-01
-7.22637951e-01 -2.52075940e-01 -5.00735343e-01 7.80849576e-01
-1.79576659e+00 -9.43155050e-01 -3.58535379e-01 6.73485816e-01
4.87903774e-01 -4.03697789e-01 8.90389383e-01 -4.63643163e-01
-6.90347433e-01 4.22626048e-01 -9.51626971e-02 -1.09369844e-01
3.38981032e-01 -1.45111406e+00 3.58404517e-01 4.42438930e-01
-5.87287307e-01 5.31965435e-01 4.62671548e-01 -1.11803257e+00
-1.59730470e+00 -1.14673519e+00 -8.58714990e-03 -3.95074904e-01
6.44799411e-01 -1.74941823e-01 -6.63268924e-01 7.23830104e-01
1.31700203e-01 -1.57034189e-01 2.94519663e-01 2.36180574e-01
-8.28563645e-02 3.78666371e-01 -1.09880948e+00 9.35587347e-01
1.05221069e+00 -1.20240532e-01 -7.39876270e-01 2.82334164e-02
8.50897253e-01 -1.00927126e+00 -7.66823649e-01 1.86662450e-01
3.71090978e-01 -9.80014205e-01 1.01640737e+00 -4.61429566e-01
3.13890547e-01 -6.13871336e-01 1.82126593e-02 -1.56946921e+00
-4.81560677e-01 -8.50689769e-01 -4.56869364e-01 6.26471877e-01
5.10932744e-01 -6.77660108e-01 1.00634015e+00 5.97266853e-01
-2.85938680e-01 -9.99206662e-01 -1.02114284e+00 -9.53823864e-01
4.77121100e-02 -5.60015738e-02 8.77111256e-01 5.64970553e-01
3.43957782e-01 2.70903200e-01 -8.19651112e-02 3.49456728e-01
7.04444945e-01 1.70154408e-01 7.50198662e-01 -1.06449080e+00
-4.74639088e-01 -5.57619452e-01 4.25953865e-01 -9.73785937e-01
2.78607726e-01 -8.64866793e-01 2.82856971e-01 -1.82633507e+00
-1.40197918e-01 -9.63329971e-01 -3.37022513e-01 1.06451714e+00
1.74160928e-01 -4.92347449e-01 4.58155423e-01 2.73993134e-01
-1.28214085e+00 8.17902923e-01 1.31055093e+00 -4.38267812e-02
-1.16688240e+00 -3.84068489e-01 -5.55200338e-01 8.46577227e-01
9.15429294e-01 -4.72137779e-01 -5.56486428e-01 -4.15980309e-01
3.59625041e-01 2.69257784e-01 4.99655426e-01 -1.10727155e+00
5.31067491e-01 -7.04669297e-01 2.21461505e-01 -5.85507333e-01
5.35116792e-01 -6.54790580e-01 -2.83162653e-01 8.20094228e-01
-6.29527211e-01 1.88216791e-01 4.15572464e-01 4.47560638e-01
1.41714498e-01 -1.86095819e-01 6.16783321e-01 -2.93881536e-01
-9.16302204e-01 1.52313605e-01 -5.74440897e-01 2.22899929e-01
1.12875688e+00 -2.27239758e-01 -7.04111516e-01 -2.50857770e-01
-6.84093654e-01 1.30375803e+00 6.68369889e-01 2.97151566e-01
6.57948911e-01 -1.10893631e+00 -3.10359508e-01 8.17931816e-02
-5.37472218e-02 4.84867990e-01 -1.59410834e-02 7.74261475e-01
-1.86026067e-01 2.20043436e-01 -3.88325304e-01 -1.89449638e-01
-6.33556604e-01 5.68230927e-01 5.82719743e-01 -9.55843687e-01
-6.15729094e-01 6.68840647e-01 -1.22062333e-01 -7.66636372e-01
5.54620922e-01 -5.30861855e-01 -4.23304349e-01 -8.79186690e-02
3.53533655e-01 5.96909046e-01 -5.63262582e-01 3.78214725e-04
-5.56297004e-01 1.57335356e-01 -2.62285709e-01 -4.89455193e-01
1.30049741e+00 9.58673358e-02 1.03728905e-01 -4.00811294e-03
1.64776146e-01 -2.79996127e-01 -1.74914634e+00 -3.99075806e-01
1.69050410e-01 -9.54881012e-02 2.78537720e-01 -1.47669876e+00
-4.95447695e-01 2.81400561e-01 2.05930859e-01 -1.10131120e-02
8.24001253e-01 -4.81171384e-02 4.77108896e-01 9.56868052e-01
7.56849468e-01 -1.46785891e+00 4.10836518e-01 1.19751501e+00
9.38366830e-01 -1.07670105e+00 -7.76160359e-02 -6.66230544e-02
-9.90124941e-01 9.14490163e-01 1.22270620e+00 -2.19421461e-01
-9.94537175e-02 4.38781023e-01 -1.56058058e-01 -3.00648481e-01
-1.23019123e+00 -4.77976531e-01 -3.93907398e-01 9.61417556e-01
-3.45385671e-01 -1.07453100e-01 2.42658898e-01 6.93750560e-01
-2.35969186e-01 1.21351987e-01 3.08073491e-01 1.39572442e+00
-9.28318739e-01 -1.13724411e+00 -3.94250035e-01 2.11111590e-01
2.38541022e-01 -4.50242013e-02 -1.76698789e-01 9.20344591e-01
4.28391434e-02 9.27612066e-01 -1.75670360e-03 -3.30338061e-01
2.37477407e-01 -2.43176013e-01 3.09919089e-01 -6.64400697e-01
-1.10452807e+00 3.14079970e-02 4.18322027e-01 -1.15229571e+00
1.16165362e-01 -5.10542929e-01 -2.04250026e+00 -1.71409413e-01
1.04892872e-01 4.48122889e-01 2.03760609e-01 8.09132874e-01
5.36762118e-01 8.51791620e-01 3.40822130e-01 -1.02647710e+00
-8.32883120e-01 -5.66445410e-01 -2.70736635e-01 -3.95625025e-01
3.74901026e-01 -1.06423736e+00 6.40320182e-02 -7.78935552e-01]
|
[3.894062042236328, 1.7269234657287598]
|
78d65f38-62d7-4716-b86f-e24a6cadf939
|
instructpix2pix-learning-to-follow-image
|
2211.09800
| null |
https://arxiv.org/abs/2211.09800v2
|
https://arxiv.org/pdf/2211.09800v2.pdf
|
InstructPix2Pix: Learning to Follow Image Editing Instructions
|
We propose a method for editing images from human instructions: given an input image and a written instruction that tells the model what to do, our model follows these instructions to edit the image. To obtain training data for this problem, we combine the knowledge of two large pretrained models -- a language model (GPT-3) and a text-to-image model (Stable Diffusion) -- to generate a large dataset of image editing examples. Our conditional diffusion model, InstructPix2Pix, is trained on our generated data, and generalizes to real images and user-written instructions at inference time. Since it performs edits in the forward pass and does not require per example fine-tuning or inversion, our model edits images quickly, in a matter of seconds. We show compelling editing results for a diverse collection of input images and written instructions.
|
['Alexei A. Efros', 'Aleksander Holynski', 'Tim Brooks']
|
2022-11-17
| null |
http://openaccess.thecvf.com//content/CVPR2023/html/Brooks_InstructPix2Pix_Learning_To_Follow_Image_Editing_Instructions_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Brooks_InstructPix2Pix_Learning_To_Follow_Image_Editing_Instructions_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['text-based-image-editing']
|
['computer-vision']
|
[ 8.05705667e-01 3.95107418e-01 1.82150602e-01 -7.13628531e-01
-4.59358811e-01 -6.64715230e-01 8.14128160e-01 -9.26050022e-02
-5.84790051e-01 5.16990244e-01 -1.54440468e-02 -5.82028985e-01
4.00020182e-01 -7.18010604e-01 -1.25242949e+00 -1.79727495e-01
3.67173195e-01 6.18305027e-01 2.18645498e-01 8.09222162e-02
7.13362575e-01 2.49033153e-01 -1.33567178e+00 6.53464139e-01
1.05378842e+00 3.95937145e-01 6.31981313e-01 1.31709313e+00
-3.74782413e-01 1.42852509e+00 -6.75059259e-01 -5.37164211e-01
2.85835624e-01 -7.57928789e-01 -7.38228261e-01 4.08288211e-01
7.96797931e-01 -7.55282342e-01 -2.81982601e-01 9.18807209e-01
1.65923521e-01 1.54289827e-01 9.29755509e-01 -8.85813415e-01
-1.36878765e+00 6.27611935e-01 -5.45624197e-01 -1.19857028e-01
2.58859426e-01 5.51235378e-01 3.67690265e-01 -9.08189178e-01
1.15082765e+00 1.17837787e+00 3.17199022e-01 8.80349040e-01
-1.38589013e+00 -3.13600123e-01 1.76543713e-01 -1.94613755e-01
-1.04784310e+00 -3.56425673e-01 2.18875989e-01 -7.41687715e-01
1.19750226e+00 9.75466371e-02 4.18037206e-01 1.01085532e+00
4.55624402e-01 8.08246136e-01 1.14952374e+00 -6.46648467e-01
8.19080472e-02 3.49737823e-01 -1.88621134e-01 1.06971157e+00
-2.35178962e-01 2.67903805e-01 -3.96709800e-01 3.45450431e-01
1.15064323e+00 -1.74474925e-01 -2.21244484e-01 -2.04097942e-01
-1.26902056e+00 6.25864983e-01 1.84729785e-01 -1.64982036e-01
-2.11686179e-01 4.81464237e-01 3.13637197e-01 5.71600497e-01
3.48986447e-01 5.16942084e-01 -3.23582232e-01 -1.40236527e-01
-9.93858099e-01 6.75287321e-02 7.77674377e-01 1.21921802e+00
9.33121800e-01 2.00674105e-02 -3.12080681e-01 5.67197084e-01
-2.32021417e-02 5.71415544e-01 3.49341184e-01 -1.34748888e+00
6.10389531e-01 9.85747501e-02 9.78735015e-02 -8.41698766e-01
2.40901619e-01 1.22250728e-01 -6.34115100e-01 5.31855524e-01
4.13741529e-01 -2.03354016e-01 -1.30577934e+00 1.61502194e+00
1.32556260e-02 2.31469097e-03 -1.90328017e-01 5.95015883e-01
5.39400041e-01 9.43782449e-01 -8.36517587e-02 -1.06224827e-02
9.54585493e-01 -1.27526224e+00 -4.85997677e-01 -4.88710552e-01
7.05741405e-01 -7.81281590e-01 1.60518014e+00 5.73069930e-01
-1.46796322e+00 -8.96363616e-01 -8.73531580e-01 -6.57848835e-01
-3.62146944e-01 4.37735111e-01 1.79754212e-01 2.49330088e-01
-1.51524627e+00 7.49649644e-01 -6.80809557e-01 -4.72087622e-01
3.24159205e-01 3.33577603e-01 -3.02356035e-01 -3.12862247e-01
-4.88753736e-01 9.26535964e-01 2.25500450e-01 -1.42790779e-01
-1.24755502e+00 -8.67642045e-01 -8.36407721e-01 -2.34722663e-02
3.83755833e-01 -9.47208941e-01 1.60598481e+00 -1.40327215e+00
-1.85562515e+00 1.12886357e+00 -2.94673681e-01 -3.77438098e-01
8.67901444e-01 -2.86101401e-01 -3.33788544e-02 2.94072956e-01
-4.51274291e-02 1.27160919e+00 1.39469779e+00 -1.47731197e+00
-4.45674449e-01 -3.17261158e-03 2.55809814e-01 3.31935324e-02
-6.00956194e-02 9.73237529e-02 -9.29478407e-01 -7.91403294e-01
-6.38776779e-01 -9.53942001e-01 -2.83715725e-01 2.52212405e-01
-5.97895086e-01 3.55655968e-01 7.28744149e-01 -7.64143050e-01
9.94813323e-01 -2.19418764e+00 3.85310054e-01 3.02809000e-01
3.05790365e-01 1.47389948e-01 -6.25774086e-01 2.52988249e-01
-9.23391059e-02 1.93502307e-01 -4.72125083e-01 -6.33343875e-01
1.58900023e-02 4.75959420e-01 -5.40520787e-01 -1.47308530e-02
2.99573749e-01 1.09412837e+00 -9.31388438e-01 -6.40200734e-01
2.16816604e-01 4.25624043e-01 -8.73823941e-01 5.08279443e-01
-7.37426043e-01 5.51314652e-01 -6.68280125e-02 -7.50702471e-02
4.76442128e-01 -2.34870732e-01 6.76972941e-02 -5.18987812e-02
-1.18865117e-01 -8.74121487e-03 -7.84053802e-01 1.74732566e+00
-8.07942569e-01 9.03645396e-01 -7.80317113e-02 -5.57768404e-01
6.76923573e-01 -7.33329430e-02 -9.88069251e-02 -5.91032624e-01
1.99350435e-03 -2.35688081e-03 -2.71102607e-01 -7.30287671e-01
6.37851417e-01 1.53792560e-01 1.37538135e-01 1.12623477e+00
2.07564607e-01 -7.82187998e-01 6.02548301e-01 7.98040509e-01
1.00579059e+00 5.52635133e-01 2.63271909e-02 -1.14802070e-01
2.60586262e-01 3.40458900e-02 -3.92157398e-03 1.02900016e+00
3.06360424e-01 7.11008847e-01 5.48505127e-01 -4.69498605e-01
-1.24240792e+00 -1.03962708e+00 5.11175692e-01 1.22016132e+00
-2.50436962e-01 -4.80680317e-01 -1.22925663e+00 -7.99015582e-01
-1.56316265e-01 1.07445657e+00 -8.23256493e-01 -7.88898170e-02
-5.91571271e-01 -8.37765727e-03 2.88820595e-01 4.84284133e-01
4.68778074e-01 -1.21558833e+00 -6.79793775e-01 -2.54193991e-02
2.08426863e-01 -1.01885724e+00 -1.35578609e+00 6.36294484e-02
-7.68418729e-01 -9.31069016e-01 -5.82013726e-01 -9.91205513e-01
1.56578231e+00 9.10883956e-03 1.39321947e+00 3.83581817e-01
-3.72814685e-01 7.65822470e-01 5.32806925e-02 -3.85859072e-01
-1.06428313e+00 -1.84736237e-01 -3.00054550e-01 -9.49220639e-03
-1.95213575e-02 -3.66059393e-01 -3.95046771e-01 1.67132452e-01
-1.25155294e+00 6.11944377e-01 7.83101559e-01 6.54374242e-01
6.18028224e-01 -1.30537778e-01 -1.56293400e-02 -1.64270329e+00
7.19536722e-01 1.13000557e-01 -8.91077161e-01 3.75578433e-01
-4.38942552e-01 4.23261464e-01 8.26175511e-01 -4.79548872e-01
-1.45486510e+00 4.79609579e-01 1.58591464e-01 -4.83309358e-01
-1.81338042e-01 4.05090928e-01 2.43383914e-01 -1.43402651e-01
8.35350633e-01 4.54015315e-01 4.17331867e-02 -2.15199366e-01
9.55391169e-01 3.11878234e-01 1.02550745e+00 -7.32879102e-01
8.75358284e-01 2.60938197e-01 -4.27826881e-01 -6.79977179e-01
-7.17068255e-01 3.43711734e-01 -9.74994838e-01 -1.96997240e-01
1.07483065e+00 -6.16241992e-01 -3.65726322e-01 4.37195778e-01
-1.39636278e+00 -1.24823964e+00 -5.02706528e-01 9.24093053e-02
-8.03728998e-01 2.54092366e-01 -9.46344852e-01 -2.31882080e-01
4.75214049e-02 -1.42226088e+00 9.19739962e-01 1.11260183e-01
-4.18320447e-01 -1.14591026e+00 -1.38862610e-01 9.66039747e-02
5.35820246e-01 -4.55607940e-03 1.10676408e+00 -1.04314141e-01
-1.09716713e+00 4.13810760e-02 -3.13607097e-01 5.98541081e-01
1.18507780e-01 5.72702289e-01 -6.88847303e-01 -2.57523395e-02
-1.43655956e-01 -5.53707123e-01 9.06295061e-01 2.28710040e-01
1.70012033e+00 -4.87690181e-01 -2.25408331e-01 6.36763573e-01
1.24732339e+00 1.22723684e-01 8.16574872e-01 -1.47323608e-01
7.63630033e-01 3.93757761e-01 3.52218688e-01 1.68234929e-01
4.35005635e-01 2.74331897e-01 1.66266888e-01 -1.41172528e-01
-3.72993678e-01 -4.88516748e-01 6.35911942e-01 1.01728964e+00
-9.37888473e-02 -2.08217561e-01 -7.06365526e-01 5.00096858e-01
-1.55882502e+00 -9.75216746e-01 -5.38416440e-03 2.20397210e+00
1.28559685e+00 2.31553823e-01 -2.52138704e-01 -4.71390426e-01
4.89788681e-01 -2.27059852e-02 -7.47654617e-01 -7.97267973e-01
2.04361260e-01 4.75454271e-01 3.39897633e-01 1.05938709e+00
-7.33369112e-01 1.13832116e+00 7.19614744e+00 5.40563107e-01
-1.17792642e+00 -6.69308007e-02 8.14136863e-01 -1.89224988e-01
-5.17064095e-01 3.74896340e-02 -5.55192232e-01 3.66046131e-01
9.41067219e-01 -2.86696255e-01 8.89333487e-01 5.72707236e-01
2.92653710e-01 -2.70843476e-01 -1.51638877e+00 7.43205190e-01
4.27832454e-01 -1.56321585e+00 5.08184612e-01 -5.60290106e-02
1.09226942e+00 -1.53184384e-01 3.04821521e-01 3.07622463e-01
1.03907073e+00 -1.02897143e+00 8.02481949e-01 6.78457022e-01
1.07208848e+00 -4.22086656e-01 4.85037453e-02 4.63251531e-01
-6.45200789e-01 1.10599712e-01 -3.07714045e-01 5.81034757e-02
1.34575039e-01 5.70856690e-01 -7.66420186e-01 -7.94466436e-02
3.25407296e-01 7.92756855e-01 -7.33340025e-01 5.27316809e-01
-6.87928021e-01 4.16044831e-01 -1.01941917e-02 2.57020921e-01
1.59050539e-01 -3.88026983e-01 -4.20065783e-02 1.42559671e+00
5.23817241e-01 2.52774775e-01 9.64340121e-02 1.17168880e+00
-3.75140727e-01 -1.51292458e-01 -8.16562414e-01 -3.46010149e-01
1.99769795e-01 1.20782626e+00 -6.20379865e-01 -8.31054628e-01
-4.28535193e-01 1.75988770e+00 4.05751258e-01 6.33518219e-01
-9.54790652e-01 -6.89875126e-01 2.07795322e-01 1.07773133e-01
5.05141854e-01 -4.97850299e-01 -1.13036312e-01 -1.21072447e+00
-2.24033296e-01 -1.07948411e+00 -1.19066097e-01 -1.42584538e+00
-8.42649281e-01 5.24356186e-01 -5.92477620e-02 -7.38815069e-01
-3.83225352e-01 -7.54956901e-01 -7.90872812e-01 7.68244267e-01
-1.32487035e+00 -8.97372901e-01 -4.06687409e-01 5.35300374e-01
7.37200737e-01 6.04978837e-02 5.90018034e-01 2.00730771e-01
-4.58697289e-01 4.98937279e-01 -3.82743701e-02 1.67422295e-01
9.87718940e-01 -1.52340961e+00 8.15376759e-01 7.50062764e-01
2.92644113e-01 8.82417619e-01 8.25738728e-01 -6.87368572e-01
-1.32463539e+00 -1.21821642e+00 7.64348567e-01 -8.42194319e-01
5.04236281e-01 -4.30685371e-01 -8.00425529e-01 1.22146118e+00
8.79061997e-01 -2.37906173e-01 4.59302098e-01 -5.43314397e-01
-3.51564020e-01 2.18049794e-01 -7.84464836e-01 9.22647595e-01
1.12726390e+00 -8.76765907e-01 -4.99978244e-01 5.26695549e-01
7.72110879e-01 -8.85974824e-01 -6.73532605e-01 -4.26932156e-01
2.48439267e-01 -8.88218760e-01 7.41001308e-01 -6.44456506e-01
1.05889440e+00 -3.64057302e-01 2.73329794e-01 -1.70374966e+00
-1.02685504e-01 -8.23759317e-01 1.17040358e-01 1.04027414e+00
6.04056656e-01 -2.70888895e-01 4.46298569e-01 8.47312033e-01
-2.26396739e-01 -4.37890887e-01 -3.77710685e-02 -5.45287907e-01
7.05936030e-02 -3.66136611e-01 3.96995634e-01 7.52585292e-01
-1.92136452e-01 4.09476757e-01 -3.48702639e-01 -2.24615082e-01
4.80885565e-01 2.77304426e-02 1.25347137e+00 -6.72226369e-01
-6.28810167e-01 -2.62985647e-01 2.57146358e-01 -1.56812465e+00
2.18808457e-01 -8.98073018e-01 2.79199749e-01 -1.40500963e+00
2.69460320e-01 -1.18505031e-01 3.50196302e-01 7.21210003e-01
-2.42723897e-01 2.18820110e-01 3.23133916e-01 8.52948427e-02
-5.26084423e-01 3.06584746e-01 1.54355848e+00 -3.48068297e-01
-1.08550645e-01 -3.94401848e-01 -6.11805320e-01 6.60424888e-01
4.60512161e-01 -4.19781029e-01 -8.53807211e-01 -1.16822076e+00
4.33756351e-01 -1.44212861e-02 2.17323124e-01 -7.87053287e-01
3.33741277e-01 -3.59799623e-01 3.88011605e-01 -2.14690819e-01
2.75751688e-02 -5.41749597e-01 4.31277715e-02 2.94573486e-01
-8.33957970e-01 3.47889364e-01 2.38149062e-01 5.12223661e-01
-1.22061640e-01 -3.30454975e-01 7.65249014e-01 -4.85176086e-01
-7.54688919e-01 3.10246736e-01 -5.67770660e-01 8.45345575e-03
9.97643113e-01 -2.90624142e-01 -5.57124376e-01 -7.47138858e-01
-9.05597627e-01 2.03484580e-01 8.54458511e-01 2.48503476e-01
6.72732711e-01 -1.08251739e+00 -5.17053664e-01 3.77423316e-01
-8.00780356e-02 -4.55894098e-02 1.29292652e-01 5.19193530e-01
-9.48488653e-01 1.66116253e-01 -1.78201124e-01 -5.67581415e-01
-1.14179862e+00 6.81943893e-01 2.87912160e-01 -1.95763990e-01
-5.42430043e-01 9.95104015e-01 6.25549078e-01 -5.91774404e-01
3.84581015e-02 -7.89758384e-01 4.53188121e-01 -4.23598677e-01
7.18597651e-01 -2.72624314e-01 -2.59361655e-01 -1.53961241e-01
3.21821034e-01 7.16705084e-01 -4.07212794e-01 -4.07748401e-01
1.15728116e+00 -1.87639877e-01 -2.71896958e-01 2.88365752e-01
1.06503606e+00 1.56764716e-01 -1.86550081e+00 -8.75166655e-02
-2.60874361e-01 -5.97998738e-01 -2.18073800e-01 -9.59859252e-01
-1.12555873e+00 1.26193583e+00 1.34324729e-01 -8.36674273e-02
1.00970852e+00 -2.90616274e-01 5.47034025e-01 7.37953424e-01
1.14911020e-01 -1.23708129e+00 6.24517024e-01 7.20102429e-01
1.08024311e+00 -1.16315269e+00 -1.10801250e-01 -2.48515904e-01
-9.21258271e-01 1.11323261e+00 8.64234805e-01 -9.75469947e-02
3.91129136e-01 5.54184496e-01 3.01856637e-01 -4.73275557e-02
-1.03397977e+00 3.98215085e-01 -7.52034932e-02 7.35629559e-01
3.23159724e-01 -2.56696135e-01 3.16499949e-01 -1.65530946e-02
-1.42007679e-01 6.03369236e-01 8.69696975e-01 8.74143004e-01
-3.49097043e-01 -1.20795846e+00 -1.29436225e-01 5.44954538e-01
3.71325836e-02 -3.30956370e-01 -6.45830750e-01 6.28417432e-01
2.15521269e-03 4.48359996e-01 2.50119418e-01 -1.86624140e-01
2.10458502e-01 3.94819416e-02 8.51919472e-01 -9.34800744e-01
-6.53376222e-01 -3.12442482e-01 -9.57454145e-02 -7.09402978e-01
-3.61041933e-01 -2.98595726e-01 -1.36132765e+00 -4.63627219e-01
6.71554655e-02 -2.38376725e-02 5.51041245e-01 8.90781701e-01
6.74403489e-01 5.62536716e-01 2.28284523e-01 -1.10246396e+00
-3.86945665e-01 -6.97035968e-01 -2.98960835e-01 5.34052312e-01
2.09620029e-01 3.00881593e-03 -2.06211153e-02 1.02463698e+00]
|
[11.312725067138672, -0.19570960104465485]
|
9696f7ca-507a-42b8-ad64-8accacdd2392
|
federated-distillation-based-indoor
|
2205.11440
| null |
https://arxiv.org/abs/2205.11440v2
|
https://arxiv.org/pdf/2205.11440v2.pdf
|
Federated Distillation based Indoor Localization for IoT Networks
|
Federated distillation (FD) paradigm has been recently proposed as a promising alternative to federated learning (FL) especially in wireless sensor networks with limited communication resources. However, all state-of-the art FD algorithms are designed for only classification tasks and less attention has been given to regression tasks. In this work, we propose an FD framework that properly operates on regression learning problems. Afterwards, we present a use-case implementation by proposing an indoor localization system that shows a good trade-off communication load vs. accuracy compared to federated learning (FL) based indoor localization. With our proposed framework, we reduce the number of transmitted bits by up to 98%. Moreover, we show that the proposed framework is much more scalable than FL, thus more likely to cope with the expansion of wireless networks.
|
['El Mehdi Amhoud', 'Marwa Chafii', 'Yaya Etiabi']
|
2022-05-23
| null | null | null | null |
['indoor-localization']
|
['computer-vision']
|
[ 7.86368083e-03 4.30620760e-01 -1.59583598e-01 -3.13889205e-01
-8.10050011e-01 -2.41482899e-01 6.34648561e-01 2.62702674e-01
-6.04443550e-01 1.25642896e+00 -5.68123311e-02 -5.40315151e-01
-4.49666470e-01 -1.11185288e+00 -7.43695199e-01 -8.15075397e-01
-5.86839497e-01 1.12733305e-01 9.64818448e-02 -6.21025227e-02
-2.57248104e-01 4.11689818e-01 -1.48014820e+00 -3.06031942e-01
8.98894608e-01 1.28181100e+00 2.07796171e-01 5.67699909e-01
-1.35507330e-01 7.84064710e-01 -6.26396656e-01 -4.50079143e-01
3.31145555e-01 -2.31709689e-01 -7.05374062e-01 -5.51212668e-01
3.97020608e-01 -1.00910328e-01 -2.79311508e-01 7.86777794e-01
8.05370033e-01 -3.82686332e-02 1.82131693e-01 -1.34819424e+00
8.94933194e-03 1.01688504e+00 -1.04611665e-02 -1.02264449e-01
3.51316720e-01 -4.04424399e-01 5.72674513e-01 -4.65074211e-01
5.77248275e-01 7.71000504e-01 7.19995856e-01 5.43025434e-01
-1.10756946e+00 -8.07856560e-01 2.45095178e-01 2.82231271e-01
-1.08968616e+00 -5.10490537e-01 7.94002533e-01 1.54692248e-01
5.70855558e-01 4.75421578e-01 5.26785314e-01 1.22060823e+00
1.27341390e-01 7.13920474e-01 1.04860425e+00 -6.06101394e-01
7.98851728e-01 2.95515805e-01 -2.60453552e-01 5.75850129e-01
7.26032972e-01 -6.93233088e-02 -4.36556935e-01 -1.07370235e-01
1.69384331e-01 1.36601403e-01 -2.53466815e-01 -7.47153699e-01
-1.04131579e+00 6.39727354e-01 7.45580971e-01 7.17393160e-01
-4.05922532e-01 4.60585713e-01 4.71900851e-01 4.03357536e-01
7.22777128e-01 1.94812298e-01 -6.58573389e-01 -1.19339488e-01
-1.22403479e+00 -3.91160557e-03 1.21205330e+00 8.67903531e-01
9.62903440e-01 -4.73536439e-02 5.88395707e-02 4.84927982e-01
2.46109039e-01 5.41036904e-01 4.95288014e-01 -9.41502750e-01
4.32511300e-01 3.80738556e-01 8.69476348e-02 -9.08234179e-01
-6.98468626e-01 -7.31034517e-01 -1.02272677e+00 -6.57306053e-03
2.59738117e-01 -6.54758155e-01 -1.23365819e-01 1.75566328e+00
3.27072412e-01 2.62093246e-01 2.75354266e-01 3.68346155e-01
3.27043951e-01 2.75333345e-01 -4.84589078e-02 -3.42835069e-01
7.84418702e-01 -9.58342373e-01 -7.62906313e-01 6.31856769e-02
7.75893450e-01 -3.07674259e-01 4.93454158e-01 6.34872258e-01
-8.65135670e-01 -3.59058201e-01 -1.25951850e+00 5.54083169e-01
-6.67692900e-01 -6.03372604e-02 8.82947505e-01 1.35198832e+00
-1.16012371e+00 6.69928789e-01 -9.81341958e-01 -7.74969637e-01
4.63126540e-01 8.20333719e-01 -4.15041476e-01 -2.85901219e-01
-9.15938377e-01 7.31481373e-01 1.57434866e-01 -3.21114957e-01
-6.12637818e-01 -4.03641164e-01 -7.88228095e-01 6.55102581e-02
1.83082670e-01 -7.08357096e-01 1.17861962e+00 -6.17932022e-01
-1.62783718e+00 1.78867951e-01 3.41391899e-02 -9.94229198e-01
6.98141575e-01 -1.95401490e-01 -6.74303651e-01 -3.79742123e-02
-1.22186102e-01 2.66625673e-01 6.04745209e-01 -1.23123562e+00
-8.54359329e-01 -5.27533114e-01 4.03795838e-02 -4.25521761e-01
-9.92438138e-01 -5.11960089e-01 1.04467897e-02 -3.63830209e-01
2.06075273e-02 -6.19366229e-01 -4.96319383e-01 2.48584569e-01
3.35864685e-02 -6.75843985e-05 1.27184093e+00 -9.55698043e-02
1.22324812e+00 -2.01606297e+00 -1.96543828e-01 3.06745201e-01
1.20542757e-01 2.14431152e-01 -1.91942714e-02 5.39593697e-01
4.72763717e-01 -9.86117125e-02 -7.44899213e-02 -8.00992072e-01
1.74502671e-01 4.38497603e-01 8.89160782e-02 8.63110065e-01
-5.23783982e-01 4.37986612e-01 -1.07892311e+00 -5.05359173e-01
4.29428577e-01 5.33920765e-01 -5.30493677e-01 2.12553948e-01
-2.13993683e-01 6.10183895e-01 -5.12179434e-01 6.15985811e-01
9.23892498e-01 -1.07562542e-02 4.70514983e-01 -2.89224125e-02
-3.98824453e-01 3.15355025e-02 -1.15227997e+00 2.31000233e+00
-1.17128265e+00 3.79173577e-01 5.03251910e-01 -1.15642107e+00
1.22402179e+00 4.31478918e-01 9.12947655e-01 -8.02575350e-01
1.52460128e-01 6.10631406e-01 -5.16239882e-01 -3.99145275e-01
2.91132718e-01 2.55659163e-01 -1.21728756e-01 4.23648179e-01
4.21434343e-01 3.24141055e-01 8.83036703e-02 -4.38309014e-02
1.61181474e+00 2.79844463e-01 4.41615909e-01 -2.72766590e-01
8.83826673e-01 -2.44863093e-01 2.70910889e-01 8.69646013e-01
-2.29854211e-01 -2.40605930e-03 -1.76062778e-01 -3.74059528e-01
-5.31883061e-01 -6.03048027e-01 -1.66796431e-01 9.05420840e-01
4.03375387e-01 -6.17773175e-01 -7.02203035e-01 -1.01893091e+00
1.16116956e-01 5.51408470e-01 -4.70095277e-01 4.05856483e-02
-5.02753735e-01 -4.29488480e-01 8.89341414e-01 9.61919054e-02
9.04115617e-01 -6.94982767e-01 -1.08457923e+00 3.39573354e-01
1.16130680e-01 -9.35563922e-01 3.48728836e-01 7.71977186e-01
-7.45473921e-01 -1.08898604e+00 -6.73953891e-01 -5.61864972e-01
4.43225950e-01 3.31556678e-01 8.36111307e-01 -1.16197050e-01
-7.08318129e-02 4.92553234e-01 -6.61933601e-01 -2.30625972e-01
-2.06311882e-01 4.83542055e-01 8.11557919e-02 5.14062680e-02
2.27657147e-02 -6.97816133e-01 -7.65706241e-01 2.17221513e-01
-7.63171017e-01 -3.73921454e-01 4.69425023e-01 6.85791731e-01
1.20493263e-01 1.93100378e-01 6.67451978e-01 -1.06032610e+00
2.37195313e-01 -9.18378294e-01 -5.46497226e-01 2.62636870e-01
-8.26914728e-01 3.30828160e-01 1.21827972e+00 -2.50676215e-01
-1.02350390e+00 2.39342570e-01 -4.67966229e-01 -2.97216009e-02
-2.90998131e-01 1.30441412e-01 -2.40298852e-01 -7.27264404e-01
6.61749184e-01 -1.27806082e-01 -1.72292262e-01 -7.02553988e-01
3.80567431e-01 9.08183575e-01 3.08460295e-01 -6.34790838e-01
7.36691713e-01 6.62173510e-01 3.05804223e-01 -6.41795754e-01
-4.79448259e-01 -2.85553545e-01 -3.99497807e-01 -6.94883540e-02
1.61943048e-01 -7.43965566e-01 -8.25525820e-01 -3.43477540e-02
-7.82517612e-01 -3.75919402e-01 -5.44082463e-01 5.86890042e-01
-6.68442070e-01 2.34633997e-01 1.01339154e-01 -1.05073512e+00
-5.94834626e-01 -7.85123706e-01 7.77730882e-01 5.62417917e-02
1.45715162e-01 -1.14304626e+00 2.35691980e-01 -1.71704844e-01
1.14169383e+00 4.92359638e-01 5.81120074e-01 -7.68320620e-01
-2.84719378e-01 -3.49092960e-01 4.83914278e-02 -1.14907343e-02
2.97617942e-01 -4.80791509e-01 -1.14102602e+00 -7.17972875e-01
-9.00982618e-02 -2.22959384e-01 6.90095186e-01 5.74876145e-02
1.23891747e+00 -5.87957263e-01 -6.01164520e-01 9.54675436e-01
1.65623105e+00 -1.60175040e-01 3.13767731e-01 6.05711758e-01
7.93778151e-02 3.62239748e-01 8.08959782e-01 8.83542061e-01
4.19444531e-01 6.91957712e-01 1.06655228e+00 -2.50795454e-01
-2.46009752e-01 -3.10540050e-01 3.37737679e-01 4.19349760e-01
3.50365669e-01 -4.17936981e-01 -5.72086036e-01 5.05975842e-01
-2.01706576e+00 -6.62845612e-01 1.00670986e-01 2.37667370e+00
3.19579601e-01 1.48639232e-02 1.55647531e-01 6.55967176e-01
3.44544917e-01 2.79994160e-01 -3.13778907e-01 -2.78275192e-01
4.52416800e-02 5.37299395e-01 9.80610728e-01 2.06834838e-01
-1.00111556e+00 3.99756819e-01 5.78624201e+00 5.83471596e-01
-1.34357154e+00 5.78460574e-01 3.57361548e-02 1.37333393e-01
-1.98410690e-01 -1.42495111e-01 -4.76723284e-01 4.91559297e-01
1.18420017e+00 9.45289880e-02 2.92435944e-01 1.14635944e+00
-7.98486173e-02 1.49152666e-01 -9.78322864e-01 9.80168402e-01
-1.55673057e-01 -1.25590491e+00 -4.82400954e-01 8.25209469e-02
6.02840543e-01 2.99780250e-01 -2.07893252e-01 3.35863650e-01
3.72786641e-01 -6.96246624e-01 7.25124240e-01 2.60518581e-01
8.81816208e-01 -7.57909417e-01 9.97425735e-01 6.53613269e-01
-1.26104009e+00 -5.60077846e-01 -2.63825893e-01 -9.80561450e-02
-2.27466747e-01 6.81237459e-01 -7.51580834e-01 9.24175918e-01
8.25144351e-01 5.23353100e-01 -5.89363217e-01 1.32697630e+00
1.95738114e-02 4.94237721e-01 -5.85167646e-01 -2.63856143e-01
9.86654684e-02 4.12139744e-01 2.51733512e-01 1.31079972e+00
7.31655300e-01 -4.54348862e-01 1.29352868e-01 3.56989622e-01
-9.10250619e-02 2.20443264e-01 -7.65613377e-01 6.28141522e-01
6.71975017e-01 1.40983653e+00 -4.87375587e-01 9.76261273e-02
-5.20232022e-01 8.86838138e-01 2.23692656e-01 4.19009663e-02
-5.14218986e-01 -3.80549788e-01 4.33906376e-01 2.87912991e-02
4.16763663e-01 -2.58247018e-01 -3.54392752e-02 -9.55600441e-01
-3.13978791e-01 -5.68949640e-01 3.53004396e-01 -5.83169274e-02
-1.12064064e+00 7.95851529e-01 -2.80670106e-01 -1.09737647e+00
-3.58440638e-01 -4.25296634e-01 -2.92841524e-01 3.05643141e-01
-1.86985636e+00 -1.24370587e+00 -7.81280518e-01 9.68120933e-01
2.72515118e-01 -6.10423088e-02 1.44741821e+00 5.86092710e-01
-4.60002154e-01 7.24429786e-01 4.96449500e-01 -3.83636892e-01
6.14637136e-01 -1.23349535e+00 4.06471677e-02 8.46894264e-01
9.54576358e-02 4.56655562e-01 7.49678910e-01 -1.07552573e-01
-1.81096816e+00 -1.30106878e+00 6.10949814e-01 1.33996814e-01
4.05194908e-01 -5.49383104e-01 -9.74253118e-02 2.64516354e-01
1.85218483e-01 7.82933295e-01 7.02930927e-01 7.69847259e-02
-2.25947216e-01 -8.97812068e-01 -1.80851722e+00 8.24684277e-02
8.38984311e-01 -7.48938620e-02 2.78588831e-01 3.21017146e-01
5.95662117e-01 -1.54128000e-01 -9.74423289e-01 5.03045380e-01
5.91647029e-01 -1.05673552e+00 7.05549836e-01 2.52890319e-01
-4.45482314e-01 -2.51714200e-01 -6.18773520e-01 -1.18599427e+00
-3.27782966e-02 -8.10057938e-01 -6.87495649e-01 1.27842712e+00
5.07848710e-02 -7.91800916e-01 1.01817775e+00 6.17140047e-02
1.39634788e-01 -5.80439091e-01 -1.37480950e+00 -8.35378587e-01
-2.85718918e-01 -6.29020691e-01 9.27310944e-01 5.60235023e-01
7.73538798e-02 -1.73658952e-01 -4.49444890e-01 4.03407514e-02
7.65839517e-01 -1.20273836e-01 1.06927466e+00 -1.42765999e+00
-3.97218943e-01 1.18429177e-02 -4.96108353e-01 -1.00698984e+00
8.71795490e-02 -7.47315645e-01 -1.21392086e-01 -1.43101311e+00
-4.88423288e-01 -1.17312050e+00 -6.11816108e-01 6.68852150e-01
5.91638029e-01 4.60996032e-01 1.77780241e-01 2.77387872e-02
-1.01027155e+00 3.80940080e-01 5.87989628e-01 -2.48091072e-01
-3.46019603e-02 5.33085406e-01 -5.27899981e-01 3.51574808e-01
9.86616671e-01 -4.97415304e-01 -2.65366167e-01 -3.54483068e-01
1.97785795e-02 1.29481286e-01 1.50935516e-01 -1.58967400e+00
5.67215085e-01 2.62135118e-01 3.16810906e-01 -3.13827991e-01
1.52417988e-01 -1.66458571e+00 2.96485722e-01 8.48575294e-01
-1.00715756e-01 -2.60792136e-01 2.37465799e-02 5.58403313e-01
3.42979236e-03 -8.62398148e-02 4.90029037e-01 4.23558243e-02
-4.24793839e-01 1.84440345e-01 -1.23439975e-01 -6.09731555e-01
1.32164431e+00 -1.74938917e-01 -4.38524157e-01 -3.90750080e-01
-4.73294407e-01 1.19631872e-01 4.12962973e-01 1.77554861e-01
3.47361982e-01 -9.18377101e-01 -3.66106778e-01 2.88478047e-01
7.50024151e-03 -8.29307139e-02 -1.59266025e-01 8.25138271e-01
-4.78122711e-01 7.25380480e-01 -3.42804939e-02 -3.29031795e-01
-1.05685687e+00 5.47962010e-01 2.72494495e-01 -3.26286703e-01
-6.45338595e-01 6.77617967e-01 -9.03640926e-01 -6.13273859e-01
8.92553568e-01 -1.20433085e-01 4.07857597e-02 -1.43935710e-01
2.17797309e-01 7.84397602e-01 5.42526126e-01 -3.36952835e-01
-6.46591663e-01 2.56973416e-01 4.70119953e-01 3.12326342e-01
1.62984300e+00 -1.66965187e-01 -1.04932018e-01 7.01879561e-02
1.21051002e+00 2.48418525e-01 -1.12676775e+00 -3.08702558e-01
2.46812999e-01 -3.61043364e-01 3.02594215e-01 -8.76010239e-01
-1.17861569e+00 4.84158069e-01 1.10778069e+00 4.64593381e-01
1.36518133e+00 -1.77385405e-01 7.92635024e-01 5.49333334e-01
1.30577493e+00 -5.16430080e-01 -2.86084712e-01 4.15611453e-02
5.94788417e-02 -1.09143972e+00 6.71997070e-02 -1.84476942e-01
2.55980700e-01 1.26234889e+00 3.02739680e-01 -1.50409877e-01
8.30656111e-01 6.83966637e-01 -2.63564110e-01 1.08650893e-01
-6.65478468e-01 -2.47479275e-01 -4.34351981e-01 9.38749611e-01
3.52664769e-01 -9.99206677e-02 -5.43096483e-01 7.00637817e-01
-1.52378544e-01 1.84143305e-01 1.69174850e-01 1.03919780e+00
-4.74291831e-01 -1.57706952e+00 -4.85996485e-01 1.38664663e-01
-5.30878663e-01 3.77266943e-01 -8.64789239e-04 6.46790683e-01
7.12892413e-01 1.31839395e+00 -3.19687963e-01 -4.50392753e-01
6.92174807e-02 -3.18717718e-01 6.48978055e-01 -3.34254235e-01
-7.83630013e-01 -3.35570812e-01 -3.68451998e-02 -7.51453698e-01
-8.04921985e-01 -2.86058098e-01 -1.13230753e+00 -3.00275564e-01
-5.39518058e-01 5.25667250e-01 1.38250828e+00 7.58195996e-01
3.08167577e-01 5.51391900e-01 9.59099114e-01 -7.71727860e-01
-5.11473060e-01 -6.82725132e-01 -6.67197704e-01 -2.00007603e-01
6.85856402e-01 -5.26919901e-01 -3.43534380e-01 -6.52836859e-01]
|
[6.402238845825195, 0.9739000201225281]
|
97b8fec4-615f-4e9a-9b7a-369d1797408f
|
ctran-cnn-transformer-based-network-for
|
2303.10606
| null |
https://arxiv.org/abs/2303.10606v1
|
https://arxiv.org/pdf/2303.10606v1.pdf
|
CTRAN: CNN-Transformer-based Network for Natural Language Understanding
|
Intent-detection and slot-filling are the two main tasks in natural language understanding. In this study, we propose CTRAN, a novel encoder-decoder CNN-Transformer-based architecture for intent-detection and slot-filling. In the encoder, we use BERT, followed by several convolutional layers, and rearrange the output using window feature sequence. We use stacked Transformer encoders after the window feature sequence. For the intent-detection decoder, we utilize self-attention followed by a linear layer. In the slot-filling decoder, we introduce the aligned Transformer decoder, which utilizes a zero diagonal mask, aligning output tags with input tokens. We apply our network on ATIS and SNIPS, and surpass the current state-of-the-art in slot-filling on both datasets. Furthermore, we incorporate the language model as word embeddings, and show that this strategy yields a better result when compared to the language model as an encoder.
|
['Javad Salimi Sartakhti', 'Mehrdad Rafiepour']
|
2023-03-19
| null | null | null | null |
['intent-detection', 'slot-filling']
|
['natural-language-processing', 'natural-language-processing']
|
[ 4.19132560e-01 3.94509882e-01 -3.32562298e-01 -3.34108800e-01
-6.23785794e-01 -3.52138132e-01 5.83640337e-01 2.04771638e-01
-9.36452925e-01 2.62932628e-01 5.95663488e-01 -7.76382744e-01
6.22272253e-01 -6.53151333e-01 -6.53255403e-01 -1.49553539e-02
2.03524277e-01 4.72883731e-01 2.62502491e-01 -2.57872462e-01
2.23725900e-01 -4.22553450e-01 -1.36343539e+00 6.71689689e-01
8.73720467e-01 9.74330425e-01 4.56412941e-01 6.40265226e-01
-6.48392856e-01 1.00474715e+00 -2.08434388e-01 -4.48566318e-01
4.75672819e-02 -3.19308162e-01 -9.71963882e-01 -4.74676900e-02
4.16193269e-02 -5.84776580e-01 -5.81074893e-01 4.56192255e-01
1.97138742e-01 1.00670062e-01 3.28621209e-01 -8.88471305e-01
-7.53338158e-01 9.14563179e-01 -3.34443182e-01 -5.11991382e-02
4.56539154e-01 -1.18188560e-01 1.53586423e+00 -1.21046007e+00
5.75787961e-01 1.31432378e+00 5.41489899e-01 5.32348812e-01
-1.06728232e+00 -3.56139392e-01 3.99587631e-01 2.55870014e-01
-1.19905090e+00 -5.01649559e-01 5.53500950e-01 -3.33487958e-01
1.58748484e+00 1.57729819e-01 6.07869446e-01 9.42557991e-01
2.86933661e-01 1.37225497e+00 6.12679243e-01 -7.57632434e-01
1.87968075e-01 5.01566380e-02 4.44841862e-01 8.20154846e-01
-1.84518859e-01 -2.09632814e-02 -6.08816922e-01 1.14537098e-01
4.95008856e-01 3.11773360e-01 1.25619352e-01 -1.56629980e-01
-1.12191999e+00 1.00617039e+00 3.75207067e-01 2.64452338e-01
-2.77105808e-01 3.53006244e-01 7.20825732e-01 1.60745949e-01
6.90559387e-01 3.35685730e-01 -5.64116478e-01 -2.43910313e-01
-1.16888726e+00 8.30068365e-02 7.24044859e-01 8.77804101e-01
7.29313135e-01 -2.06950411e-01 -6.67261600e-01 9.69351947e-01
5.11781752e-01 -3.61947268e-02 7.41329253e-01 -4.60613728e-01
6.60073340e-01 7.71124423e-01 -5.93330115e-02 -2.05926955e-01
-5.21577895e-01 -4.50391144e-01 -5.26987493e-01 -3.05795878e-01
-1.13636889e-01 -2.87423562e-02 -1.53715420e+00 1.77233315e+00
1.78679880e-02 1.75194308e-01 1.50280908e-01 6.17114365e-01
6.91471875e-01 6.45071507e-01 3.95953447e-01 1.75463721e-01
1.95297325e+00 -1.49207079e+00 -8.90038252e-01 -9.06044900e-01
1.19976759e+00 -6.83066189e-01 1.17771566e+00 -2.99039017e-02
-8.73377144e-01 -4.85201031e-01 -9.65449035e-01 -7.91463435e-01
-4.72305775e-01 4.46173102e-01 6.31163001e-01 3.76009315e-01
-1.14414346e+00 4.05069590e-01 -1.00936055e+00 -5.15427589e-01
2.50288337e-01 8.12167227e-02 -3.04251641e-01 -1.07710242e-01
-1.37347949e+00 8.58081102e-01 5.73741972e-01 -1.19697452e-01
-7.71681547e-01 -7.22922325e-01 -1.52760601e+00 1.69865072e-01
3.16099584e-01 -8.44601631e-01 1.75975382e+00 -4.20937449e-01
-1.41640186e+00 9.92485821e-01 -7.13347614e-01 -8.99614036e-01
1.77899286e-01 -5.27747869e-01 -2.29364336e-01 -1.91605568e-01
3.36243898e-01 1.11030638e+00 6.55321479e-01 -7.02408612e-01
-8.19969654e-01 -9.05004069e-02 2.72956640e-01 3.82739663e-01
-3.25172842e-01 -1.86418369e-01 -7.91371882e-01 -5.64570606e-01
2.68699676e-01 -8.69793594e-01 -4.04830813e-01 -8.10021535e-02
-6.96564853e-01 -2.66877115e-01 6.60739958e-01 -9.01939631e-01
1.77145898e+00 -2.42675328e+00 -1.54582813e-01 -2.55602390e-01
2.74415523e-01 2.20417187e-01 -3.90193939e-01 7.91887820e-01
-1.44844189e-01 1.26270875e-01 -4.19636726e-01 -1.20443857e+00
2.29768395e-01 4.54146504e-01 -6.15480661e-01 6.51651770e-02
4.19693321e-01 1.20994496e+00 -6.87665939e-01 -3.20877105e-01
3.39805484e-01 2.52297044e-01 -1.09198856e+00 4.67129856e-01
-5.38336337e-01 -6.11934997e-02 -2.43047401e-01 3.75607103e-01
5.28295755e-01 -2.36735836e-01 3.27250749e-01 -5.62712215e-02
-4.14825231e-01 1.34511054e+00 -6.85883820e-01 2.09001708e+00
-9.50059712e-01 6.60889983e-01 -2.61862844e-01 -9.89658892e-01
6.65346086e-01 2.86615223e-01 2.41133526e-01 -1.21535790e+00
3.29940170e-02 2.20557973e-01 -1.75064325e-01 -1.64533406e-01
1.18736923e+00 -4.39548446e-03 -4.13020223e-01 6.51523173e-01
1.91130698e-01 9.72206220e-02 2.89897650e-01 3.43400955e-01
1.23251009e+00 1.01978801e-01 3.04309577e-01 -1.72744300e-02
2.70631313e-01 6.50637671e-02 2.76938766e-01 6.78174317e-01
1.66910172e-01 5.34963667e-01 6.40680850e-01 -4.72506911e-01
-1.16464102e+00 -9.50562239e-01 4.25933450e-02 1.43993247e+00
-1.85488109e-02 -8.35867405e-01 -7.34508991e-01 -6.70677662e-01
-7.99321309e-02 1.03221619e+00 -7.28711784e-01 -3.41649264e-01
-7.22513497e-01 -2.61321038e-01 5.51356018e-01 7.46394932e-01
5.39435446e-01 -1.04811871e+00 -7.53182113e-01 5.26495457e-01
-4.13911194e-01 -1.22429478e+00 -7.61799991e-01 8.29491079e-01
-7.55336583e-01 -6.98814690e-01 -4.39888597e-01 -1.00751007e+00
5.32837808e-01 1.51973680e-01 1.33373690e+00 2.15330556e-01
-1.34384885e-01 -1.79968476e-02 -5.40227652e-01 -5.42462021e-02
-1.35839090e-01 5.41468143e-01 -3.17371756e-01 -3.59332591e-01
4.27270204e-01 -3.02313328e-01 -6.94923401e-01 8.76288190e-02
-1.03472316e+00 5.86644828e-01 6.24002934e-01 1.01567125e+00
3.46989959e-01 -4.44514662e-01 6.45219758e-02 -9.20673907e-01
4.56002593e-01 -3.53869408e-01 -3.33556741e-01 1.34684131e-01
-5.43211401e-01 5.22324026e-01 4.83425230e-01 -2.34121438e-02
-7.68078268e-01 1.26902714e-01 -7.38742054e-01 -1.77182987e-01
5.17132021e-02 6.17731571e-01 1.10092722e-01 5.34265399e-01
2.69859642e-01 3.01726431e-01 -2.41058946e-01 -7.47798681e-01
6.60016358e-01 7.67320156e-01 4.63764429e-01 -2.23196611e-01
4.59383339e-01 3.15154046e-01 -6.39347136e-01 -4.77944165e-01
-1.22919273e+00 -6.62979782e-01 -5.37906408e-01 4.46929961e-01
9.48466063e-01 -1.11559796e+00 -4.68877405e-01 3.51302058e-01
-1.49221170e+00 -4.85066742e-01 -4.75981236e-01 1.69977561e-01
-4.74523306e-01 1.28048167e-01 -8.60974610e-01 -8.77242744e-01
-5.70491970e-01 -1.07877314e+00 1.42869425e+00 9.50852856e-02
-3.54873091e-01 -9.95620966e-01 1.85678080e-01 2.19664648e-01
4.73653555e-01 -6.09532177e-01 9.23836410e-01 -1.03988087e+00
-5.22324979e-01 4.38748598e-02 -3.60001236e-01 1.06193619e-02
-2.95297112e-02 -7.22947359e-01 -1.12057936e+00 2.54528373e-02
-3.10168475e-01 -3.88205618e-01 1.46379280e+00 2.50827879e-01
1.18958426e+00 -2.68870622e-01 -5.71304619e-01 6.05973661e-01
1.14023077e+00 2.84790426e-01 9.08388317e-01 4.76924628e-01
5.29352427e-01 3.76215130e-01 6.18657351e-01 5.67252338e-01
8.83705795e-01 5.54405034e-01 3.95612717e-01 -3.73561025e-01
-2.85971817e-02 -9.23059464e-01 4.01666313e-01 8.21619511e-01
6.16974652e-01 -5.52218676e-01 -7.70123184e-01 8.02930355e-01
-1.98682940e+00 -4.90693420e-01 2.21635506e-01 1.91021633e+00
8.39409113e-01 3.16149503e-01 -1.62091151e-01 5.81378490e-02
3.81365299e-01 5.92587888e-01 -4.41197753e-01 -7.36125767e-01
3.13766241e-01 6.10837102e-01 6.03853643e-01 7.91530132e-01
-1.32233059e+00 1.59094298e+00 6.48118973e+00 7.33755112e-01
-9.52763617e-01 2.57543683e-01 5.70479095e-01 1.52999520e-01
-7.35155821e-01 1.94448233e-01 -9.17514741e-01 2.26416111e-01
9.59736526e-01 3.06109041e-01 3.36692005e-01 9.45370138e-01
-6.49853349e-02 -2.14982629e-01 -1.21852148e+00 7.40462780e-01
-3.96222882e-02 -1.34893513e+00 -1.39567673e-01 -1.58217743e-01
2.18377486e-01 1.76678430e-02 1.17782317e-01 9.00297701e-01
3.73223543e-01 -1.01503086e+00 8.81894290e-01 9.82793197e-02
9.47943389e-01 -4.06654984e-01 6.57084227e-01 1.84420586e-01
-1.55836546e+00 -1.03823274e-01 -2.75568098e-01 -2.42949367e-01
5.64775467e-01 6.68390930e-01 -1.02408588e+00 2.94995815e-01
2.64883906e-01 8.70548785e-01 -1.79020002e-01 7.57558286e-01
-2.32028529e-01 5.97695291e-01 -2.72251517e-01 1.41958416e-01
5.60916603e-01 5.53071499e-03 2.09737062e-01 1.47226083e+00
1.88797712e-01 -2.37642959e-01 3.13671917e-01 8.39538574e-01
-1.15415610e-01 -6.24920428e-02 -3.70635599e-01 -4.58606005e-01
2.86610365e-01 1.01058054e+00 -5.89091241e-01 -5.52048624e-01
-5.39186537e-01 1.51044500e+00 4.70720440e-01 1.07641138e-01
-8.46241772e-01 -4.20699239e-01 9.83920753e-01 -1.80272218e-02
6.62268937e-01 -3.51350158e-01 -4.73320246e-01 -1.23377335e+00
-2.35308763e-02 -4.32736844e-01 4.26174372e-01 -5.55068254e-01
-6.24724329e-01 6.40466452e-01 -2.33516380e-01 -1.07312334e+00
-4.30131793e-01 -6.68219864e-01 -5.31841218e-01 8.25898170e-01
-1.91731834e+00 -1.10229063e+00 1.60925016e-01 1.96447656e-01
1.00264394e+00 1.75739452e-01 9.20879185e-01 5.90398490e-01
-5.50671637e-01 6.23060405e-01 -3.05739343e-01 3.01759720e-01
4.20329899e-01 -1.13924336e+00 1.34457397e+00 7.40519285e-01
3.31233472e-01 7.48143971e-01 4.67929125e-01 -6.66789770e-01
-1.22036004e+00 -1.19890964e+00 1.57465422e+00 -1.16835810e-01
7.51949608e-01 -9.57603633e-01 -7.72530079e-01 1.04769695e+00
4.25576001e-01 -3.31377685e-02 6.09294713e-01 2.77193248e-01
-4.93900865e-01 4.88898546e-01 -6.38701737e-01 7.21974194e-01
1.30842817e+00 -8.85047913e-01 -9.69121456e-01 1.78709134e-01
1.41783142e+00 -7.69668639e-01 -1.32085443e-01 7.47238174e-02
6.26187861e-01 -7.85490692e-01 8.70497406e-01 -5.34074128e-01
6.17365301e-01 -7.69933611e-02 -1.15664393e-01 -1.24374807e+00
-3.58815789e-01 -2.54436076e-01 -9.32398960e-02 1.00960577e+00
7.82816112e-01 -5.30812562e-01 9.04668748e-01 4.42102075e-01
-6.63853407e-01 -8.45710576e-01 -9.17413294e-01 -4.23130542e-01
-1.37180850e-01 -8.40649307e-01 5.00940621e-01 4.63762671e-01
2.68777758e-01 5.59744895e-01 -4.37114805e-01 -1.63819745e-01
9.92959440e-02 1.02417886e-01 3.53211313e-01 -8.73204887e-01
-2.14247569e-01 -1.69450507e-01 1.75910339e-01 -1.82039034e+00
3.14159572e-01 -9.43814099e-01 3.48844320e-01 -1.87779188e+00
1.09655090e-01 -4.86534208e-01 -3.86529803e-01 9.84459698e-01
-1.54008135e-01 2.28903607e-01 2.64733523e-01 -1.08433038e-01
-7.12368965e-01 8.21758091e-01 8.84844482e-01 -2.74862200e-01
-2.88664460e-01 -2.90464014e-01 -7.01703787e-01 4.73337173e-01
7.00062513e-01 -5.08039951e-01 -3.13078254e-01 -6.41071677e-01
2.23759457e-01 1.35421252e-03 1.68644503e-01 -8.10657442e-01
1.96560934e-01 2.30108127e-01 3.04494891e-02 -9.30045426e-01
5.80319583e-01 -7.94607699e-01 -4.26427752e-01 6.54899836e-01
-5.95540822e-01 1.45351395e-01 4.59711522e-01 1.89282537e-01
-1.85659125e-01 -2.40487024e-01 2.74082094e-01 -7.01288879e-02
-9.52553928e-01 1.88147500e-01 -6.90368414e-01 1.65562287e-01
4.07441616e-01 -3.83296087e-02 -3.26313645e-01 -2.14151800e-01
-5.56984603e-01 3.52388769e-01 1.86967716e-01 6.06188834e-01
6.74722016e-01 -1.29002297e+00 -2.67187595e-01 6.47245765e-01
3.77969593e-01 -8.26689079e-02 2.17609957e-01 7.49866545e-01
-2.38633588e-01 8.98814261e-01 1.18028939e-01 -4.28329378e-01
-8.50492716e-01 3.91933292e-01 1.82872817e-01 -7.89043903e-01
-6.53975010e-01 8.65809917e-01 4.55333084e-01 -7.27867901e-01
1.90571547e-01 -8.47745180e-01 -2.06383154e-01 6.99752197e-02
5.11511087e-01 -3.02640617e-01 1.44600347e-01 -2.19998002e-01
-5.91690540e-01 2.55394906e-01 -4.08936590e-01 -4.02286142e-01
1.04477310e+00 -2.21106529e-01 6.62051365e-02 3.79447490e-01
1.24891865e+00 -2.18001917e-01 -1.03380775e+00 -4.66267437e-01
1.92322195e-01 -1.55394450e-01 7.97001496e-02 -6.24703467e-01
-7.55997002e-01 9.81952965e-01 3.04454237e-01 1.17078744e-01
8.99401844e-01 5.95575385e-02 1.30141985e+00 2.41239071e-01
5.00418656e-02 -1.07413518e+00 7.21905902e-02 1.22750461e+00
4.92034048e-01 -1.13766253e+00 -4.88586634e-01 -3.25433940e-01
-6.72287524e-01 8.10343266e-01 6.27739370e-01 -2.87877899e-02
4.96292353e-01 3.99480432e-01 6.00466020e-02 4.23678569e-03
-1.11486065e+00 -6.55476272e-01 2.73278236e-01 2.52690524e-01
8.93706381e-01 9.19366777e-02 -5.54576993e-01 6.09108627e-01
-4.02367800e-01 9.31002125e-02 1.65436059e-01 1.09133649e+00
-5.83729327e-01 -1.35033095e+00 -9.85217467e-03 4.60267991e-01
-2.90967196e-01 -8.17014396e-01 -1.06762432e-01 4.71381187e-01
-3.88787873e-02 8.33349884e-01 6.97832406e-01 -6.98973179e-01
2.74588019e-01 5.02552986e-01 -6.94800988e-02 -1.01852238e+00
-6.23469293e-01 -2.83516254e-02 4.26931739e-01 -7.72222281e-01
1.79818586e-01 -3.75959992e-01 -1.45988333e+00 4.27946188e-02
-1.74926698e-01 1.33963645e-01 6.78371847e-01 1.23764622e+00
5.80237389e-01 9.49892819e-01 4.05772805e-01 -5.40744126e-01
-2.07961723e-01 -1.00597012e+00 -3.01057041e-01 -8.84335581e-03
5.85276425e-01 -5.54223657e-01 -1.62313923e-01 -1.39315620e-01]
|
[12.446806907653809, 7.389571189880371]
|
bb26bd01-ecb4-4d39-8ba6-f745ab01b79b
|
neural-execution-engines-learning-to-execute
|
2006.08084
| null |
https://arxiv.org/abs/2006.08084v3
|
https://arxiv.org/pdf/2006.08084v3.pdf
|
Neural Execution Engines: Learning to Execute Subroutines
|
A significant effort has been made to train neural networks that replicate algorithmic reasoning, but they often fail to learn the abstract concepts underlying these algorithms. This is evidenced by their inability to generalize to data distributions that are outside of their restricted training sets, namely larger inputs and unseen data. We study these generalization issues at the level of numerical subroutines that comprise common algorithms like sorting, shortest paths, and minimum spanning trees. First, we observe that transformer-based sequence-to-sequence models can learn subroutines like sorting a list of numbers, but their performance rapidly degrades as the length of lists grows beyond those found in the training set. We demonstrate that this is due to attention weights that lose fidelity with longer sequences, particularly when the input numbers are numerically similar. To address the issue, we propose a learned conditional masking mechanism, which enables the model to strongly generalize far outside of its training range with near-perfect accuracy on a variety of algorithms. Second, to generalize to unseen data, we show that encoding numbers with a binary representation leads to embeddings with rich structure once trained on downstream tasks like addition or multiplication. This allows the embedding to handle missing data by faithfully interpolating numbers not seen during training.
|
['Danai Koutra', 'Yujun Yan', 'Kevin Swersky', 'Parthasarathy Ranganathan', 'Milad Hashemi']
|
2020-06-15
| null |
http://proceedings.neurips.cc/paper/2020/hash/c8b9abffb45bf79a630fb613dcd23449-Abstract.html
|
http://proceedings.neurips.cc/paper/2020/file/c8b9abffb45bf79a630fb613dcd23449-Paper.pdf
|
neurips-2020-12
|
['learning-to-execute']
|
['computer-code']
|
[ 4.66993302e-01 -1.70066312e-01 -1.64293796e-01 -3.14394265e-01
-3.73626560e-01 -9.39115942e-01 4.50005561e-01 5.39926648e-01
-4.94168282e-01 6.10038280e-01 2.26527572e-01 -7.75175631e-01
1.29222438e-01 -1.13517189e+00 -1.03721428e+00 -3.68417591e-01
-1.67941168e-01 3.59230846e-01 4.36482951e-02 -3.03691208e-01
3.51801187e-01 5.87133408e-01 -1.70947647e+00 5.47871649e-01
7.38151252e-01 6.59496427e-01 -6.63495511e-02 9.35805023e-01
-3.26640934e-01 7.71578431e-01 -6.83714688e-01 -4.69548881e-01
3.92022610e-01 -3.72655600e-01 -7.98538089e-01 -3.01897675e-01
6.51509821e-01 -6.00139141e-01 -5.23320138e-01 9.93621111e-01
8.99776723e-03 1.99893355e-01 7.97674298e-01 -1.36808467e+00
-1.25047815e+00 7.64069855e-01 -7.77966827e-02 2.83882886e-01
4.17919368e-01 3.37425292e-01 1.22113883e+00 -7.80634940e-01
4.80254441e-01 1.13540876e+00 1.05557501e+00 7.46910214e-01
-1.45095098e+00 -4.21867669e-01 -3.59242670e-02 -5.94821200e-02
-1.13440549e+00 -2.08839014e-01 2.72704899e-01 -4.49970961e-01
1.17105663e+00 1.76101431e-01 4.10090834e-01 7.38422096e-01
1.07638381e-01 5.10089159e-01 6.80936992e-01 -4.27236915e-01
2.73697257e-01 3.93084763e-03 2.40674064e-01 6.60982192e-01
5.09109139e-01 6.77237958e-02 -3.62435430e-01 -9.25270766e-02
7.10101247e-01 2.00642481e-01 -4.17237371e-01 -2.39214376e-01
-1.29223943e+00 7.75098264e-01 6.31695449e-01 4.45321620e-01
-1.15233362e-02 4.31704253e-01 4.94696826e-01 6.63190842e-01
-1.08288653e-01 7.02991843e-01 -6.53766036e-01 -8.60645920e-02
-8.22714746e-01 2.39982828e-01 8.77498984e-01 9.44137156e-01
8.16183388e-01 6.09048195e-02 8.64234865e-02 3.71442437e-01
-3.06355536e-01 5.92686087e-02 7.12211549e-01 -8.49565566e-01
6.60761297e-01 6.20558739e-01 5.53556159e-02 -8.69663775e-01
-3.59411865e-01 -6.00348353e-01 -7.56209433e-01 2.61550605e-01
9.03523386e-01 6.95012361e-02 -9.74023342e-01 2.13180208e+00
-2.48090729e-01 -2.34318212e-01 2.97882050e-01 6.91402376e-01
3.58883262e-01 9.30195630e-01 1.56315967e-01 1.82889640e-01
1.10805559e+00 -6.62547648e-01 -2.51807481e-01 -5.23897469e-01
1.25474441e+00 -2.62043595e-01 1.47148979e+00 2.23115563e-01
-1.32239115e+00 -6.54693961e-01 -1.43303847e+00 -5.29944956e-01
-7.08207905e-01 -2.25241154e-01 7.29586840e-01 6.25507951e-01
-1.28109920e+00 1.18607163e+00 -6.28215313e-01 -3.17897081e-01
4.85015601e-01 3.99900943e-01 -3.23674947e-01 -2.24697024e-01
-1.17676806e+00 9.31649804e-01 6.12872541e-01 -7.10593462e-02
-5.35226226e-01 -1.04823279e+00 -9.82488632e-01 6.07586265e-01
-8.70726183e-02 -7.69503474e-01 1.34386015e+00 -1.14867961e+00
-8.81126165e-01 7.55716443e-01 -1.39529750e-01 -6.34190679e-01
2.42937192e-01 1.56447832e-02 -2.69496053e-01 -8.82618409e-03
-1.42965287e-01 7.24406779e-01 6.41311407e-01 -1.02658129e+00
-2.43314072e-01 -4.19594139e-01 2.03398377e-01 3.93096590e-03
-4.71987695e-01 -3.43096286e-01 3.06074202e-01 -7.38161743e-01
7.94079304e-02 -6.56382620e-01 -2.00610965e-01 3.35526913e-01
-1.02937900e-01 -1.03364058e-01 4.11650628e-01 -3.75537455e-01
1.11982620e+00 -2.28727651e+00 8.71272981e-02 1.41773105e-01
1.99600130e-01 2.27768600e-01 -2.03140035e-01 6.57734513e-01
-3.62848729e-01 2.84414530e-01 -4.90330517e-01 9.32368487e-02
2.75907129e-01 3.82765800e-01 -5.74527204e-01 3.54801059e-01
5.23637354e-01 1.04786813e+00 -9.26656306e-01 2.75341440e-02
-1.96313560e-01 6.41189923e-04 -9.46781337e-01 2.04460453e-02
-4.32085931e-01 -5.57333231e-02 1.19026126e-02 2.21452788e-01
5.48904538e-01 -5.52665293e-01 1.55408010e-01 -9.09557845e-03
1.28853023e-01 6.12209499e-01 -1.05381191e+00 1.75314093e+00
-6.99959755e-01 5.95539629e-01 -1.55929625e-01 -1.08821464e+00
5.50370932e-01 1.00059725e-01 -2.17621341e-01 -4.64084804e-01
1.50122875e-02 4.97997940e-01 3.94938707e-01 -2.20727757e-01
5.74268997e-01 -5.44954956e-01 -1.98179260e-01 7.22399652e-01
-4.60317694e-02 -1.73630416e-01 3.10015321e-01 3.72368217e-01
1.32860184e+00 -1.56325206e-01 2.07073092e-01 -6.28151596e-02
4.45238858e-01 8.61363932e-02 2.09173992e-01 9.26721990e-01
6.20051771e-02 4.33934212e-01 5.61399043e-01 -5.06938756e-01
-1.38704598e+00 -1.37721145e+00 -8.55952576e-02 1.39503372e+00
-1.30125925e-01 -4.72462565e-01 -4.92494375e-01 -4.78922278e-01
2.51582175e-01 8.94911289e-01 -6.99209571e-01 -5.47307432e-01
-6.85040951e-01 -4.79106337e-01 6.13076329e-01 1.18462515e+00
1.65794060e-01 -9.92769182e-01 -7.21891344e-01 3.56653720e-01
4.00150776e-01 -9.58727598e-01 -5.02432585e-01 7.06633508e-01
-1.27180946e+00 -9.40776289e-01 -5.62415898e-01 -1.11675549e+00
9.10687923e-01 -2.90724963e-01 1.20257735e+00 4.70133275e-01
-2.18292579e-01 2.53064513e-01 1.20208245e-02 8.25147703e-02
-6.98626161e-01 4.36879583e-02 -6.09371513e-02 -5.03078640e-01
4.92198467e-01 -7.86071301e-01 -3.82346481e-01 2.93103117e-03
-1.20432639e+00 -1.24004565e-01 6.05433643e-01 9.68767405e-01
-8.43887776e-02 -3.06470245e-02 5.37956297e-01 -6.97527885e-01
7.30787754e-01 -5.01497149e-01 -3.61069709e-01 2.19241064e-02
-2.96367675e-01 6.88877285e-01 1.27840984e+00 -6.52020514e-01
-4.31154132e-01 -2.60923684e-01 -1.22389175e-01 -1.82069480e-01
-1.58875436e-01 4.10506845e-01 2.24749465e-03 1.67858154e-01
8.88618350e-01 3.79990369e-01 3.87072712e-02 -2.53183722e-01
5.83144963e-01 5.08461893e-01 8.77813399e-01 -8.73341382e-01
8.83309245e-01 2.79871851e-01 6.65397942e-02 -4.47856307e-01
-6.98674917e-01 -3.80440690e-02 -4.87697691e-01 5.00859618e-01
6.09226763e-01 -5.80620468e-01 -7.34656215e-01 1.77446395e-01
-1.24485433e+00 -5.90839207e-01 -6.67237997e-01 3.38621855e-01
-5.95420539e-01 2.25050449e-01 -1.08175516e+00 -5.08040249e-01
2.35676616e-02 -1.02681589e+00 6.67879581e-01 -1.02946550e-01
-6.30408943e-01 -1.02329075e+00 -3.18418920e-01 -4.94231135e-01
5.83796918e-01 3.45412120e-02 1.84418976e+00 -8.60018611e-01
-8.45172048e-01 -1.21799968e-01 -3.53747129e-01 6.01265013e-01
2.34589307e-03 -2.06616342e-01 -6.34620905e-01 -1.86264768e-01
-7.67048746e-02 -5.00830650e-01 1.06224537e+00 -1.31421909e-01
1.43164468e+00 -5.18161833e-01 -2.39562526e-01 7.23892450e-01
1.49168026e+00 1.83849126e-01 6.30639374e-01 3.06987077e-01
3.67577493e-01 4.29083914e-01 -2.12212801e-01 1.88725337e-01
5.72597720e-02 7.33915046e-02 3.12629640e-01 2.89396137e-01
-1.24689922e-01 -6.18714094e-01 1.57927543e-01 7.05264688e-01
4.00139034e-01 -1.30440816e-01 -9.21438038e-01 7.17591047e-01
-1.27846920e+00 -1.00108337e+00 2.89018489e-02 2.32843590e+00
1.17460036e+00 4.66723174e-01 -8.40482265e-02 6.76284492e-01
4.97400671e-01 7.10869730e-02 -6.16891980e-01 -9.40162122e-01
1.07209295e-01 6.38082623e-01 5.20180225e-01 4.08941031e-01
-6.41483724e-01 6.76332653e-01 6.87094069e+00 4.51880574e-01
-9.68280733e-01 -1.54898286e-01 5.60085475e-01 2.00770900e-01
-6.51537120e-01 -8.95785093e-02 -4.66357648e-01 4.50492680e-01
1.31324255e+00 -8.53200704e-02 6.50050938e-01 6.98451400e-01
-5.42219937e-01 1.63737237e-01 -1.93783581e+00 7.86197424e-01
-7.71217868e-02 -1.57213664e+00 3.19143206e-01 -2.67629057e-01
5.81558108e-01 -3.73556316e-01 3.69849324e-01 5.32460928e-01
4.43042278e-01 -1.66857719e+00 6.65570378e-01 5.36698587e-02
9.50692773e-01 -6.00304484e-01 3.20212215e-01 3.75768274e-01
-8.64976346e-01 -4.37076896e-01 -4.73814815e-01 -6.32275343e-01
-1.86538622e-01 2.69780219e-01 -7.71732152e-01 -7.14683980e-02
2.49807894e-01 6.14293098e-01 -4.78232414e-01 9.32496011e-01
-1.73746407e-01 2.79664546e-01 -3.03755194e-01 -2.20115885e-01
3.87679845e-01 4.75207977e-02 8.05228129e-02 1.13684821e+00
5.95066965e-01 1.34057522e-01 -3.12272906e-01 1.08822834e+00
-4.19720978e-01 -3.02790761e-01 -7.93025017e-01 -3.44400793e-01
5.77081800e-01 5.43693185e-01 -5.52521706e-01 -4.66130078e-01
-5.51453888e-01 1.01538098e+00 5.45575619e-01 6.08617604e-01
-6.87657773e-01 -6.06063902e-01 6.55044436e-01 1.72964498e-01
4.95597541e-01 -5.32672167e-01 -6.74773633e-01 -1.02511919e+00
3.53605822e-02 -9.42675352e-01 3.58020008e-01 -8.15140128e-01
-1.17688680e+00 3.37387949e-01 -1.28163576e-01 -8.78039122e-01
-3.19115192e-01 -9.38039482e-01 -6.27991140e-01 9.50254738e-01
-1.18215597e+00 -4.35651153e-01 -1.63562025e-03 3.09789896e-01
2.50094771e-01 5.43098003e-02 7.92344272e-01 1.90713719e-01
-6.04682118e-02 8.19151878e-01 1.50468498e-01 3.41956198e-01
3.07025373e-01 -1.27107513e+00 6.86968863e-01 6.95306599e-01
2.28080243e-01 1.14652765e+00 6.30042613e-01 -2.09786966e-01
-1.50592887e+00 -1.07633162e+00 9.52239037e-01 -5.58382571e-01
7.61159241e-01 -4.78898019e-01 -1.31189775e+00 1.03977120e+00
-9.41306651e-02 2.68556744e-01 6.38720334e-01 6.03538901e-02
-1.04016888e+00 4.71833088e-02 -1.09094095e+00 7.68076837e-01
1.29228151e+00 -8.17889392e-01 -9.86848474e-01 1.79092363e-01
8.97886276e-01 -3.74301702e-01 -6.00110173e-01 2.22755119e-01
4.59924638e-01 -8.87443244e-01 1.05959439e+00 -1.09309089e+00
8.98546815e-01 -1.86574548e-01 -3.44146430e-01 -1.30925000e+00
-2.18557835e-01 -4.71948206e-01 -2.44650155e-01 7.62474954e-01
5.99954367e-01 -7.84876525e-01 8.73659790e-01 3.86926472e-01
-1.13948323e-01 -6.87727869e-01 -6.48800194e-01 -9.47345138e-01
6.32646739e-01 -5.11036098e-01 8.63672912e-01 8.40565145e-01
1.46779209e-01 2.43862152e-01 2.56852984e-01 1.03120677e-01
2.42342651e-01 2.91265547e-01 3.65746349e-01 -1.02855647e+00
-4.65910316e-01 -7.03772962e-01 -6.13223195e-01 -1.44426823e+00
1.29219204e-01 -1.37077248e+00 -4.27479632e-02 -1.34188604e+00
-5.41062690e-02 -6.05273902e-01 -1.65000543e-01 4.36248273e-01
4.90910709e-02 2.08730310e-01 1.69396341e-01 4.16737013e-02
-1.06668361e-01 4.30829525e-01 1.09954488e+00 -1.18948340e-01
9.56300125e-02 -2.17227936e-01 -9.62437570e-01 6.07022345e-01
7.29322135e-01 -5.08450568e-01 -3.13079208e-01 -6.84562564e-01
5.74890852e-01 5.90505525e-02 6.34253979e-01 -1.22756636e+00
2.32010514e-01 6.66082054e-02 5.95064163e-01 -3.66100103e-01
1.51696458e-01 -7.93896019e-01 -1.90354526e-01 6.56084538e-01
-8.44048142e-01 5.63559115e-01 4.31770682e-01 5.02478719e-01
6.95568025e-02 -5.44523954e-01 6.68289781e-01 -1.57344088e-01
-5.30395687e-01 -3.17523140e-03 -3.71402174e-01 4.09179211e-01
6.58602476e-01 -3.70423853e-01 -3.55720580e-01 -3.01282912e-01
-7.31970906e-01 -3.74348871e-02 8.41864526e-01 6.98427260e-02
5.55296659e-01 -1.27970052e+00 -4.47006166e-01 4.34032977e-01
-3.95973921e-02 1.62958559e-02 -1.40496671e-01 4.28045452e-01
-7.27253258e-01 4.49128717e-01 -2.22038060e-01 -3.04131061e-01
-6.09088063e-01 9.80894625e-01 3.24322343e-01 -4.93788049e-02
-5.29759765e-01 8.85440111e-01 2.27576956e-01 -5.12960017e-01
3.69255871e-01 -1.17356777e+00 4.88657594e-01 -2.58612692e-01
6.30238295e-01 -6.81177294e-03 1.97116539e-01 9.65103414e-03
-3.13879073e-01 4.69696999e-01 -4.14753817e-02 1.64832950e-01
1.16569948e+00 3.04778427e-01 -7.07312077e-02 4.68030423e-01
1.56215048e+00 -9.53851268e-02 -1.12145078e+00 -1.59249246e-01
5.90754189e-02 -3.55114251e-01 -5.31775057e-01 -6.66675627e-01
-6.54717922e-01 1.24021053e+00 1.12598918e-01 2.67306417e-01
9.49643254e-01 -1.76119015e-01 1.05592906e+00 6.29822075e-01
2.52069920e-01 -6.11173987e-01 1.44602045e-01 7.67016649e-01
6.25025928e-01 -8.78965676e-01 -2.65399843e-01 -1.03725374e-01
-1.16661318e-01 1.48354411e+00 4.21102136e-01 -2.65579194e-01
2.87150800e-01 6.20732605e-01 -3.39082807e-01 1.63105294e-01
-8.63889575e-01 -7.56975496e-03 3.46497074e-03 6.65512562e-01
4.40129399e-01 -2.08921507e-01 -3.48049635e-03 5.12800276e-01
-5.42485297e-01 4.32693809e-02 8.50971401e-01 1.18053854e+00
-5.26292622e-01 -9.27993238e-01 -3.62623155e-01 6.33756459e-01
-2.81561136e-01 -5.26226640e-01 3.70251723e-02 9.44268584e-01
1.72537323e-02 5.54584324e-01 4.99959439e-01 -2.24827126e-01
4.38956991e-02 5.03812253e-01 7.66685784e-01 -8.46201599e-01
-5.52408755e-01 -9.76922572e-01 -1.10632136e-01 -2.78499782e-01
1.30858600e-01 -3.65578890e-01 -1.55552030e+00 -6.01163208e-01
2.28989959e-01 2.18107894e-01 3.28508705e-01 7.84312248e-01
2.17940509e-01 5.54182172e-01 1.36896759e-01 -6.37083471e-01
-1.22477388e+00 -4.58517343e-01 -4.12446052e-01 6.19692326e-01
7.93740749e-01 -1.83797702e-01 -6.11419559e-01 -8.65484849e-02]
|
[9.480319023132324, 7.181978225708008]
|
cb7a3382-700a-4db1-96d8-e3d0e93b1fe0
|
deep-dynamic-scene-deblurring-from-optical
|
2301.07329
| null |
https://arxiv.org/abs/2301.07329v1
|
https://arxiv.org/pdf/2301.07329v1.pdf
|
Deep Dynamic Scene Deblurring from Optical Flow
|
Deblurring can not only provide visually more pleasant pictures and make photography more convenient, but also can improve the performance of objection detection as well as tracking. However, removing dynamic scene blur from images is a non-trivial task as it is difficult to model the non-uniform blur mathematically. Several methods first use single or multiple images to estimate optical flow (which is treated as an approximation of blur kernels) and then adopt non-blind deblurring algorithms to reconstruct the sharp images. However, these methods cannot be trained in an end-to-end manner and are usually computationally expensive. In this paper, we explore optical flow to remove dynamic scene blur by using the multi-scale spatially variant recurrent neural network (RNN). We utilize FlowNets to estimate optical flow from two consecutive images in different scales. The estimated optical flow provides the RNN weights in different scales so that the weights can better help RNNs to remove blur in the feature spaces. Finally, we develop a convolutional neural network (CNN) to restore the sharp images from the deblurred features. Both quantitative and qualitative evaluations on the benchmark datasets demonstrate that the proposed method performs favorably against state-of-the-art algorithms in terms of accuracy, speed, and model size.
|
['Jimmy Ren', 'Jianbo Liu', 'Furong Zhao', 'Xing Wei', 'Shangchen Zhou', 'Daoye Wang', 'Jinshan Pan', 'Jiawei Zhang']
|
2023-01-18
| null | null | null | null |
['deblurring']
|
['computer-vision']
|
[ 2.29765326e-02 -8.79355788e-01 1.85103431e-01 -1.81065515e-01
-2.12179706e-01 -4.46061760e-01 2.83262640e-01 -8.05612445e-01
-3.65426481e-01 7.26034164e-01 6.47940755e-01 -2.95550078e-02
-7.78180137e-02 -2.80533671e-01 -6.51743352e-01 -7.56228626e-01
3.74497563e-01 -4.63645488e-01 2.36590996e-01 6.56831861e-02
4.43797231e-01 4.95986581e-01 -1.43473077e+00 2.35887002e-02
1.10170019e+00 9.46172953e-01 4.41288710e-01 7.70347297e-01
1.99926421e-01 1.10989332e+00 -5.70168018e-01 -2.74384413e-02
1.88591644e-01 -4.51706797e-01 -6.07385874e-01 2.07590818e-01
7.86371768e-01 -1.03727973e+00 -7.94683158e-01 1.46737051e+00
4.46990192e-01 3.55275184e-01 3.15916806e-01 -6.91871881e-01
-1.49485052e+00 1.54344976e-01 -9.02084291e-01 5.26317477e-01
1.81200340e-01 4.28466231e-01 4.65883017e-01 -9.32657957e-01
2.95047820e-01 1.31190169e+00 4.96794939e-01 5.74056864e-01
-1.05804133e+00 -6.95366979e-01 -5.89743666e-02 6.14853323e-01
-1.24722576e+00 -6.90215111e-01 6.79492831e-01 -2.46997640e-01
3.88507843e-01 2.94557303e-01 4.99587238e-01 9.40418005e-01
2.71384060e-01 7.64007151e-01 1.06323695e+00 -2.47722790e-02
-1.41382650e-01 -2.69524008e-01 1.73608989e-01 6.06198907e-01
4.38232780e-01 2.96344310e-01 -2.40838334e-01 3.50440681e-01
1.20317852e+00 4.67753708e-01 -1.08881485e+00 1.62218753e-02
-1.29926193e+00 4.23050672e-01 8.85839641e-01 1.83789477e-01
-4.33162093e-01 2.94859558e-01 2.21907839e-01 3.25245014e-03
5.71497381e-01 4.21434194e-01 -2.43265450e-01 -1.52596131e-01
-9.33413148e-01 -3.01812240e-03 2.64323354e-01 5.47149777e-01
6.76762998e-01 2.99237520e-01 -3.08399558e-01 1.00920761e+00
2.96760947e-01 5.46266615e-01 7.05738068e-01 -1.13457894e+00
2.00660199e-01 2.41894498e-01 5.81954658e-01 -1.18952274e+00
-1.14245236e-01 -4.21830505e-01 -1.30786586e+00 3.11340034e-01
3.72899264e-01 -1.73647761e-01 -9.23752010e-01 1.55806279e+00
7.49449804e-02 7.82584310e-01 -1.64814778e-02 1.74290204e+00
7.13662744e-01 8.30300570e-01 -2.50753582e-01 -3.58191073e-01
1.24300420e+00 -1.38393331e+00 -1.03051984e+00 -4.22767550e-01
-4.87932656e-03 -1.04422462e+00 7.96004951e-01 2.69872189e-01
-1.14485741e+00 -8.54722619e-01 -8.94935429e-01 -4.55890238e-01
6.82833940e-02 3.96450669e-01 4.49425906e-01 2.46988967e-01
-1.10037136e+00 5.54827034e-01 -7.53149450e-01 -9.08411965e-02
4.57143575e-01 -5.63308820e-02 -1.85790658e-01 -4.27040786e-01
-1.24964643e+00 1.00051022e+00 -1.20042386e-02 7.63790309e-01
-9.33295906e-01 -5.36375761e-01 -8.95815492e-01 1.36114120e-01
1.02630906e-01 -8.81724596e-01 1.00826561e+00 -1.08671832e+00
-1.51444316e+00 2.94936359e-01 -4.18394119e-01 -2.39051938e-01
5.27896285e-01 -7.22804248e-01 -3.44079375e-01 1.66013494e-01
-6.03893325e-02 4.89784718e-01 1.28345335e+00 -1.11854994e+00
-5.25616050e-01 3.50349173e-02 -5.03838882e-02 3.30372155e-01
-2.72299618e-01 1.42796278e-01 -5.38947403e-01 -8.48762870e-01
-5.04342616e-02 -6.63971901e-01 -7.74313360e-02 2.11439356e-01
-2.76807904e-01 1.98483542e-01 1.11251366e+00 -1.06066167e+00
1.22313511e+00 -2.21981430e+00 1.56425133e-01 -6.47339463e-01
3.94607365e-01 6.88067794e-01 -2.01277509e-01 -2.50180781e-01
-1.51071638e-01 -1.42944410e-01 -1.76064834e-01 -2.12530211e-01
-4.49067622e-01 -4.40673195e-02 -4.91115302e-01 6.80594444e-01
1.37516353e-02 9.43204522e-01 -1.06338370e+00 -7.75040612e-02
6.71698749e-01 7.69771755e-01 -3.57793987e-01 4.58813369e-01
1.42241225e-01 4.64633495e-01 -1.59657717e-01 3.03320259e-01
1.12164879e+00 -3.68929386e-01 -4.53025997e-01 -5.56755066e-01
-2.48949319e-01 -1.02313846e-01 -1.07040036e+00 1.62866879e+00
-4.85126436e-01 1.04156995e+00 8.46179500e-02 -4.91739333e-01
6.45860255e-01 -3.60143883e-03 1.39765143e-01 -4.18142498e-01
2.41589785e-01 6.70256093e-02 -6.85266033e-02 -7.95785666e-01
6.56626225e-01 7.33410046e-02 4.87885743e-01 3.64615798e-01
-3.38492930e-01 5.69203608e-02 -2.49603596e-02 -1.08912438e-01
7.79133737e-01 7.73320124e-02 3.16642560e-02 -1.27905458e-02
7.29648948e-01 -6.32683814e-01 6.22839510e-01 5.39884448e-01
-4.40723807e-01 1.02748382e+00 -5.35687916e-02 -5.75766444e-01
-8.39122295e-01 -1.04151571e+00 8.54619220e-02 6.29788756e-01
7.04029679e-01 1.62224472e-01 -7.44439542e-01 -4.66472834e-01
-1.54862806e-01 4.72320080e-01 -4.99092966e-01 -4.70323324e-01
-4.66254085e-01 -6.46618068e-01 1.85782284e-01 3.71998847e-01
1.04557240e+00 -9.63565886e-01 -4.91835207e-01 6.52002320e-02
-5.10370374e-01 -1.16235900e+00 -1.27404201e+00 -6.45357549e-01
-7.22603917e-01 -9.92623925e-01 -1.29755330e+00 -8.86460006e-01
8.43665063e-01 1.13302052e+00 5.04084647e-01 1.24574274e-01
-1.58665508e-01 -8.88215080e-02 -1.33051261e-01 1.14668995e-01
-1.38466671e-01 -2.66472667e-01 8.42611492e-02 4.54707444e-01
7.24482909e-02 -4.27248597e-01 -1.16405582e+00 3.79092604e-01
-1.17490196e+00 2.51235664e-01 6.20605350e-01 8.97264779e-01
7.75096053e-03 1.26506686e-01 3.06057334e-01 -2.92250901e-01
9.45126235e-01 -9.76187661e-02 -5.80088973e-01 1.61385998e-01
-5.58230579e-01 1.57241058e-03 6.97909534e-01 -6.44710183e-01
-1.47305989e+00 -1.79578483e-01 2.35559762e-01 -1.07352102e+00
-6.41221704e-04 2.28778243e-01 1.65143400e-01 -2.17065245e-01
5.53909779e-01 4.35999691e-01 3.84426713e-02 -5.14027357e-01
4.76858109e-01 8.55198801e-01 7.25655258e-01 7.76932910e-02
8.10146093e-01 5.18463969e-01 -2.69205898e-01 -6.76000297e-01
-1.09416103e+00 -4.01663244e-01 -2.85135925e-01 -3.08055639e-01
8.35446596e-01 -1.02927279e+00 -7.43849874e-01 1.15201926e+00
-1.47831726e+00 -5.15388511e-02 2.06499308e-01 7.11187065e-01
-1.53924108e-01 7.35632241e-01 -1.07523382e+00 -5.88944614e-01
-5.06951988e-01 -1.29309833e+00 8.17606747e-01 8.36544991e-01
3.03050101e-01 -8.40630293e-01 -1.57614723e-01 3.47995669e-01
8.71319890e-01 -1.67725205e-01 2.58813679e-01 4.09985065e-01
-8.77042651e-01 1.00590917e-03 -9.40361381e-01 7.14821517e-01
6.00416362e-01 -5.16499616e-02 -1.06524348e+00 -2.98497319e-01
2.94676661e-01 -5.59988944e-03 1.15638649e+00 7.65647948e-01
1.26494861e+00 -5.11485279e-01 -4.47444199e-03 9.30079460e-01
1.29985034e+00 -4.36574891e-02 9.98835623e-01 2.47855350e-01
1.08526945e+00 2.77552575e-01 3.36676806e-01 1.13070995e-01
2.68952489e-01 5.44117928e-01 3.14585209e-01 -3.34012479e-01
-6.11448586e-01 -1.20950798e-02 4.84082490e-01 7.65945017e-01
-1.68917075e-01 4.02052887e-04 -3.92445773e-01 6.05748057e-01
-1.94057512e+00 -1.10073090e+00 -2.50008076e-01 2.05141568e+00
1.05132806e+00 -1.70445234e-01 -3.38931590e-01 -3.95319879e-01
1.10404634e+00 4.38207060e-01 -6.66579187e-01 9.38518159e-03
-1.96303055e-01 -1.53867334e-01 5.55989444e-01 7.43470788e-01
-1.03482091e+00 9.76135373e-01 5.61579132e+00 5.31036496e-01
-1.46059489e+00 1.49031922e-01 7.25487709e-01 -1.87612966e-01
1.50084734e-01 -1.06112964e-01 -3.28048080e-01 8.46289456e-01
6.04519069e-01 -6.17330931e-02 1.07478690e+00 6.18798912e-01
7.11538851e-01 -1.06759734e-01 -6.61406994e-01 1.40325856e+00
1.71822518e-01 -1.34797907e+00 -2.43771635e-02 -2.84710675e-01
8.95777464e-01 1.25723943e-01 1.59576684e-01 -4.57771607e-02
2.50257403e-01 -9.20660436e-01 5.83171070e-01 1.05508733e+00
7.96693206e-01 -4.41777319e-01 7.28059173e-01 1.73649386e-01
-9.74758804e-01 -8.50987658e-02 -5.99297464e-01 -1.38465464e-01
3.44361752e-01 7.38296688e-01 -1.74076721e-01 3.35622072e-01
9.22987759e-01 1.25980723e+00 -6.20811403e-01 1.51555574e+00
-3.58899862e-01 3.46923769e-01 2.12762266e-01 1.77043334e-01
6.62030876e-02 -4.41583484e-01 5.70526719e-01 1.12998772e+00
4.23726261e-01 1.37789389e-02 -3.24344903e-01 1.02006018e+00
-2.82720685e-01 -4.72500205e-01 -2.10124284e-01 2.91645437e-01
2.40525231e-01 1.37092316e+00 -3.04863900e-01 -2.99024761e-01
-3.94512713e-01 1.45769465e+00 1.30499735e-01 8.15508604e-01
-9.30414855e-01 -6.48083568e-01 1.02030325e+00 -2.34984815e-01
2.94182062e-01 -2.56963968e-01 -2.83080693e-02 -1.75489461e+00
3.41204479e-02 -7.32132196e-01 -4.52581868e-02 -1.49278629e+00
-1.40773129e+00 5.89492142e-01 -4.19948936e-01 -1.37778151e+00
1.21495105e-01 -4.65026289e-01 -5.38399458e-01 1.18719804e+00
-1.98976254e+00 -7.71320760e-01 -8.67260396e-01 4.99845833e-01
7.03713357e-01 4.44674551e-01 2.35705942e-01 4.11177993e-01
-8.68271649e-01 1.52805105e-01 3.90761673e-01 2.21292570e-01
1.16485095e+00 -9.86510038e-01 3.96737009e-01 1.36737108e+00
-3.03591341e-01 7.66028941e-01 7.53807127e-01 -4.51158047e-01
-1.22683203e+00 -1.23068464e+00 4.12904233e-01 -1.56768769e-01
5.56902230e-01 4.93035391e-02 -1.16658008e+00 4.33869928e-01
4.58342344e-01 5.69040418e-01 -2.04353392e-01 -5.12742400e-01
-2.60555327e-01 -4.39542800e-01 -7.89043546e-01 6.19944751e-01
8.55182171e-01 -6.30513310e-01 -5.78787982e-01 5.52396327e-02
8.34330320e-01 -6.31600201e-01 -4.19882417e-01 1.58035949e-01
4.79155779e-01 -1.10778451e+00 1.07715130e+00 -2.67903030e-01
6.30960882e-01 -7.09717810e-01 3.70139986e-01 -1.65477383e+00
-5.91027677e-01 -7.95024037e-01 -2.78133780e-01 1.16133428e+00
-9.03825313e-02 -8.28863561e-01 3.01769018e-01 6.01764202e-01
-6.38290644e-02 -4.65247542e-01 -5.18627286e-01 -4.73188639e-01
-4.08432961e-01 -2.62402594e-02 4.58790958e-01 9.29276168e-01
-5.16364753e-01 3.70008975e-01 -8.98123860e-01 3.08184057e-01
6.90720975e-01 1.74989477e-01 5.30906558e-01 -8.83311391e-01
-5.76364063e-02 -5.48891306e-01 -2.62273669e-01 -1.64477253e+00
8.29798356e-02 -2.52432555e-01 2.30339527e-01 -1.59976149e+00
2.89580226e-01 1.21058682e-02 -3.79748851e-01 8.83236602e-02
-6.56363666e-01 2.40089640e-01 6.99106902e-02 6.16521180e-01
-5.00318468e-01 7.07928896e-01 1.79065549e+00 -1.51345387e-01
-9.50371176e-02 -7.59698823e-02 -8.67753625e-01 7.15002239e-01
4.95701492e-01 -1.15657933e-01 -2.87493229e-01 -9.82792437e-01
-1.93366304e-01 8.39880407e-02 6.02710485e-01 -8.60289574e-01
3.49784762e-01 -3.10084432e-01 7.43125558e-01 -2.47514561e-01
1.83245301e-01 -7.03827024e-01 -3.80457789e-02 3.76268655e-01
-2.94022948e-01 -7.03246668e-02 1.29182771e-01 7.71914005e-01
-2.75829315e-01 -2.10849926e-01 1.01444352e+00 -7.13555291e-02
-6.94238484e-01 3.76083761e-01 -7.23974332e-02 -1.39430672e-01
7.52916515e-01 -6.53196424e-02 -8.17540824e-01 -6.72794878e-01
-1.52942076e-01 3.12964022e-02 5.91467142e-01 5.93457818e-01
9.01414335e-01 -1.19282830e+00 -6.81702971e-01 1.72016367e-01
-3.42726499e-01 7.39778439e-03 6.83636546e-01 9.26100552e-01
-7.56068528e-01 2.39523396e-01 -3.59393477e-01 -4.56886232e-01
-1.06256783e+00 6.74347758e-01 6.96163476e-01 1.57865718e-01
-5.15243351e-01 8.85842800e-01 5.49966097e-01 1.85008034e-01
1.55622825e-01 -4.74095315e-01 -2.87842631e-01 -3.15957993e-01
9.80680823e-01 4.47781652e-01 -2.67527014e-01 -5.92067897e-01
-1.44499436e-01 7.32340932e-01 -2.43532166e-01 1.73664242e-01
1.24605978e+00 -7.28806913e-01 -3.10013890e-01 1.43312449e-02
1.34512484e+00 -2.45048553e-02 -1.82627916e+00 -3.09307516e-01
-5.42156935e-01 -9.96379137e-01 5.06412804e-01 -7.26575017e-01
-1.55729043e+00 9.34608400e-01 7.18867838e-01 1.47810593e-01
1.45567524e+00 -4.28711057e-01 1.13195682e+00 -3.86917442e-02
-1.62295550e-01 -6.13745153e-01 2.17588916e-01 4.59168673e-01
9.49351847e-01 -1.15202332e+00 1.64538603e-02 -2.16592938e-01
-5.73127568e-01 1.33959281e+00 6.50013208e-01 -3.43975872e-01
3.43168408e-01 -1.89357370e-01 1.97536260e-01 2.07439780e-01
-4.99239892e-01 -1.31213456e-01 4.85886663e-01 2.68914819e-01
1.75081611e-01 -3.42382848e-01 3.26762460e-02 2.48675555e-01
2.44737074e-01 3.34605277e-01 8.83521914e-01 3.75985503e-01
-4.81204301e-01 -3.98108751e-01 -5.78127146e-01 2.82284737e-01
-3.94892722e-01 -2.40048364e-01 1.71470270e-02 2.30334267e-01
-1.58935618e-02 1.08186257e+00 -5.68059497e-02 -3.26814324e-01
1.82952046e-01 -5.66774607e-01 3.71860981e-01 -1.73952311e-01
-1.18356697e-01 1.43243149e-01 -3.41154993e-01 -5.27665317e-01
-5.26815355e-01 -4.31160331e-01 -8.59250009e-01 -4.72476095e-01
-5.51195562e-01 -1.63717493e-01 3.18103254e-01 8.14508975e-01
5.61192691e-01 6.72936738e-01 7.64864385e-01 -1.00572848e+00
-4.87938911e-01 -1.14665568e+00 -4.45998073e-01 5.75464249e-01
9.63866830e-01 -5.68435073e-01 -7.08667815e-01 3.22311968e-01]
|
[11.477450370788574, -2.6012463569641113]
|
f131d5c0-ade9-42ff-b725-d7ed8f0c374f
|
experimental-demonstration-of-neuromorphic
|
2112.04749
| null |
https://arxiv.org/abs/2112.04749v1
|
https://arxiv.org/pdf/2112.04749v1.pdf
|
Experimental Demonstration of Neuromorphic Network with STT MTJ Synapses
|
We present the first experimental demonstration of a neuromorphic network with magnetic tunnel junction (MTJ) synapses, which performs image recognition via vector-matrix multiplication. We also simulate a large MTJ network performing MNIST handwritten digit recognition, demonstrating that MTJ crossbars can match memristor accuracy while providing increased precision, stability, and endurance.
|
['Joseph S. Friedman', 'Sanjeev Aggarwal', 'Dimitri Houssameddine', 'Fred B. Mancoff', 'Alexander J. Edwards', 'Peng Zhou']
|
2021-12-09
| null | null | null | null |
['handwritten-digit-recognition']
|
['computer-vision']
|
[ 1.99471071e-01 -4.69834954e-01 -6.11461848e-02 1.09848261e-01
3.90112668e-01 -1.84738114e-01 2.56709158e-01 -3.53199929e-01
-6.90025210e-01 9.06059921e-01 -8.15448523e-01 -6.64674997e-01
-8.85663852e-02 -6.58921540e-01 -1.11926925e+00 -6.08229101e-01
-2.66882420e-01 2.85451412e-01 8.08935404e-01 -4.63475972e-01
6.73957944e-01 5.16737461e-01 -1.26571131e+00 3.36531490e-01
4.87114906e-01 1.44022608e+00 2.00801045e-01 5.88185608e-01
2.53213525e-01 1.14254975e+00 -9.44797397e-01 -2.32323915e-01
1.92397177e-01 1.69463471e-01 -3.65968466e-01 -7.10602283e-01
5.42527318e-01 -1.90097556e-01 -1.23014498e+00 6.07040942e-01
2.94494778e-01 8.57150927e-02 5.87423444e-01 -1.10021591e+00
-1.17174387e+00 8.97675812e-01 -1.78307235e-01 9.73471344e-01
-1.21748939e-01 3.44678015e-01 2.97497749e-01 -5.31712234e-01
3.60726237e-01 8.98750305e-01 7.83207119e-01 9.99930918e-01
-1.62916327e+00 -9.79999959e-01 -7.26680696e-01 1.71476334e-01
-1.26315928e+00 -6.34508789e-01 5.21011114e-01 1.27154723e-01
2.36729789e+00 1.09340675e-01 1.05357814e+00 1.23879528e+00
1.57783031e+00 7.04738796e-02 1.14033365e+00 4.17290181e-02
4.93000835e-01 -3.81584942e-01 4.22012389e-01 5.95364273e-01
4.77083743e-01 1.61983520e-01 -9.79476094e-01 -8.18826780e-02
1.09085870e+00 1.47634417e-01 1.89110994e-01 1.63199723e-01
-9.21815753e-01 1.47185490e-01 6.14295542e-01 2.22081140e-01
3.96823324e-03 1.29865694e+00 4.69471127e-01 7.17019618e-01
-6.43532991e-01 4.78410512e-01 3.21265340e-01 -2.49073982e-01
-6.93180919e-01 -4.72734123e-01 9.42682266e-01 7.27379918e-01
1.76563919e-01 1.01233912e+00 2.40075901e-01 8.66404414e-01
3.11981179e-02 1.36076331e+00 8.50717008e-01 -7.95110047e-01
1.25887737e-01 5.38945615e-01 -5.07361412e-01 -6.16748869e-01
-3.05381894e-01 1.23163544e-01 -9.80272532e-01 3.01431507e-01
-2.19264627e-01 3.43157381e-01 -9.68689799e-01 1.20152843e+00
-6.98702097e-01 2.35618949e-02 1.34826973e-01 7.58270144e-01
6.64399028e-01 5.36642969e-01 -3.81899700e-02 1.45587549e-01
9.70747650e-01 -4.87763405e-01 -5.54512739e-01 -3.43414962e-01
-2.07182646e-01 -1.55246198e-01 4.70580727e-01 -9.74704698e-02
-1.28468621e+00 -4.34591204e-01 -1.66045904e+00 5.74452942e-03
-4.66724664e-01 -2.51633108e-01 7.90286303e-01 7.64120936e-01
-1.64343882e+00 6.59666359e-01 -1.31591058e+00 -3.32924515e-01
1.35141879e-01 1.44092607e+00 -2.53737450e-01 5.08355796e-01
-1.03618085e+00 9.06349838e-01 2.91721463e-01 6.09734328e-03
-1.24586809e+00 -3.24767321e-01 -2.22852975e-01 7.16733560e-02
-6.95958376e-01 -7.60778725e-01 1.06456006e+00 -3.01560521e-01
-1.60617828e+00 6.09444320e-01 1.88384000e-02 -1.22987807e+00
-4.76789474e-01 8.17471564e-01 -1.04326916e+00 4.61550504e-01
-3.62698555e-01 9.64636922e-01 9.88224745e-01 -5.44874847e-01
1.50433958e-01 -4.40067142e-01 -4.18659508e-01 -6.15246177e-01
-1.00634301e+00 -3.05317134e-01 4.94362831e-01 -3.76291752e-01
4.03913647e-01 -8.92465472e-01 2.14657977e-01 -1.20598584e-01
-4.25932616e-01 4.08356376e-02 1.04633987e+00 5.72784990e-02
8.69922340e-01 -2.08061123e+00 -3.22460294e-01 4.37951684e-01
6.02323890e-01 2.05280095e-01 -2.23606691e-01 4.45135176e-01
3.21668327e-01 -1.60784498e-01 -7.30959177e-02 2.98755109e-01
-4.18319494e-01 3.27981561e-01 -8.20871353e-01 2.23010540e-01
2.62052655e-01 1.32129014e+00 -2.87331700e-01 -7.80986696e-02
-1.89183876e-01 4.40696359e-01 -1.75202131e-01 -6.06055021e-01
2.36037284e-01 6.77311942e-02 -8.58459175e-02 8.25264752e-01
4.05215114e-01 -5.18995047e-01 3.47551882e-01 -1.16834588e-01
-1.08541064e-01 4.12803650e-01 -2.42586300e-01 1.21237493e+00
-1.99612573e-01 1.22857332e+00 4.95865606e-02 -9.26736295e-01
1.43317771e+00 6.88880980e-02 3.71594399e-01 -1.63695776e+00
2.02652037e-01 7.27152109e-01 4.93833065e-01 5.65731190e-02
5.76610029e-01 4.95062806e-02 2.64254510e-01 6.18439555e-01
4.10950601e-01 -1.98568106e-01 3.74253765e-02 2.13222682e-01
1.56196320e+00 -9.83513296e-01 -7.41493523e-01 -9.27564204e-01
-5.43568879e-02 2.88318191e-02 2.16928665e-02 1.13223970e+00
-1.94506705e-01 -9.72541273e-02 -1.25012631e-02 -4.56506133e-01
-1.36912763e+00 -1.74916601e+00 -5.59131205e-01 5.40427923e-01
3.27732980e-01 1.30646199e-01 -4.15251791e-01 4.41731811e-01
3.52238923e-01 7.18103163e-03 -3.97348613e-01 -5.89678049e-01
-6.87994540e-01 -8.89319718e-01 1.47482657e+00 7.55023718e-01
8.01040351e-01 -9.93799865e-01 -9.15682912e-01 6.13715589e-01
6.15491152e-01 -1.05969298e+00 -1.60148203e-01 1.03629792e+00
-1.67827654e+00 -5.75225830e-01 5.89389205e-02 -1.32892275e+00
6.38017118e-01 3.44789475e-02 9.34881747e-01 1.00283928e-01
-7.91270614e-01 2.79442191e-01 6.17041290e-01 2.49350950e-01
-3.47728223e-01 2.82028317e-02 7.30233312e-01 -7.00692356e-01
4.46849614e-01 -1.35096610e+00 -9.42972422e-01 2.95304000e-01
-8.57974291e-01 -3.97505492e-01 4.84047562e-01 9.43356216e-01
5.42618036e-01 -1.17183715e-01 6.25200808e-01 -1.60255671e-01
8.24978113e-01 5.19681424e-02 -6.55384123e-01 8.29816833e-02
-7.72488296e-01 2.80999452e-01 8.61366034e-01 -8.74257445e-01
-3.91707778e-01 -3.27571511e-01 6.55740825e-03 -2.12878838e-01
7.88050771e-01 1.66823380e-02 7.80821562e-01 -1.13598669e+00
9.57515717e-01 9.37550008e-01 3.09364021e-01 1.05654567e-01
-4.51352954e-01 5.47736406e-01 9.48704064e-01 -4.94872719e-01
3.84008408e-01 6.72337472e-01 2.72027612e-01 -8.05716217e-01
8.51423502e-01 6.20076895e-01 -6.13646768e-02 -1.45169601e-01
1.44908085e-01 -7.49091804e-01 -1.56356263e+00 5.69381058e-01
-8.47299218e-01 -6.92573607e-01 -1.52132884e-01 1.37501985e-01
-2.00820237e-01 -2.63785303e-01 -1.80781066e+00 -5.80266416e-01
-8.56566131e-01 -9.73463714e-01 4.76090014e-01 3.32986534e-01
1.00471973e-01 -8.44740033e-01 -4.99516074e-03 -2.06082821e-01
1.18640602e+00 -4.89281595e-01 1.32172823e+00 -1.12106651e-01
-9.23855901e-01 -1.82310306e-02 -1.60080880e-01 9.46635976e-02
-3.49257499e-01 1.77473500e-02 -9.59803581e-01 -5.02566576e-01
2.10358249e-03 -6.50571704e-01 1.27946258e+00 3.57663751e-01
8.69144559e-01 -1.47772133e-01 -5.78858674e-01 3.45620483e-01
1.44703007e+00 5.86617827e-01 9.20015752e-01 1.65302545e-01
4.33499187e-01 -5.02771497e-01 -9.11393285e-01 6.72591031e-02
9.62306932e-02 3.48247796e-01 1.97801247e-01 3.34918737e-01
-2.90600032e-01 -1.75237477e-01 3.98075223e-01 1.48422575e+00
1.33647352e-01 9.28295329e-02 -1.01106536e+00 2.93491066e-01
-1.25301886e+00 -7.75464296e-01 -3.78843881e-02 1.84288979e+00
8.04060996e-01 4.07515317e-01 -3.10390800e-01 6.59326315e-02
7.08024800e-01 -3.29847187e-01 -1.30278206e+00 -7.67966986e-01
-5.78210294e-01 7.18969703e-01 1.03630722e+00 -2.40845103e-02
-3.80720973e-01 7.79672921e-01 8.63292599e+00 6.87640727e-01
-1.72244811e+00 1.88950181e-01 3.75981271e-01 -4.00066823e-01
-3.66842479e-01 -2.28337452e-01 -5.53047419e-01 6.81303918e-01
1.52332890e+00 2.49339789e-01 1.17686474e+00 2.60417432e-01
-5.10358095e-01 -2.13189825e-01 -1.03760123e+00 1.21589613e+00
-4.53143120e-01 -1.80714118e+00 3.49906117e-01 2.00899541e-01
7.94234216e-01 3.93172294e-01 6.76790953e-01 2.78675854e-01
7.94051513e-02 -1.15327835e+00 5.69963038e-01 5.81393123e-01
1.13753653e+00 -4.31752384e-01 1.99536026e-01 -1.75113857e-01
-8.31687689e-01 -7.75447428e-01 -9.55176592e-01 -3.19110930e-01
-3.75537306e-01 7.16155469e-01 -4.35581088e-01 -6.76690519e-01
8.93069983e-01 2.47335806e-01 -5.84359884e-01 8.60963106e-01
8.36996794e-01 5.49830854e-01 -4.94663715e-01 -8.10338676e-01
-8.60508159e-02 -2.48408113e-02 4.14956033e-01 7.89379358e-01
1.41924992e-01 -1.06427915e-01 -4.44146186e-01 1.44859171e+00
-7.41966963e-01 -6.67656124e-01 -8.62407207e-01 -4.91655976e-01
1.01646948e+00 8.69119227e-01 -1.14954066e+00 -3.06989610e-01
3.35586548e-01 1.11085296e+00 3.04629892e-01 2.96061546e-01
-5.21256208e-01 -7.88310349e-01 6.69465482e-01 2.21274979e-02
1.24066792e-01 -6.68192983e-01 -8.95394564e-01 -7.78626144e-01
6.27211183e-02 -2.87810326e-01 -4.41218048e-01 -6.32588625e-01
-9.80761528e-01 8.10306370e-01 -9.57711339e-01 -6.19450390e-01
1.00124851e-01 -1.01542819e+00 -5.63130915e-01 2.95677930e-01
-7.53946424e-01 -2.77512521e-01 1.50050744e-01 5.09995639e-01
-2.28049338e-01 -4.49242860e-01 7.84202039e-01 3.53007048e-01
-7.09317923e-01 8.04984927e-01 7.46991217e-01 1.01519279e-01
1.85102642e-01 -6.07028842e-01 9.50925946e-01 3.31810623e-01
-1.75116941e-01 1.13120389e+00 2.61620522e-01 -5.86231172e-01
-2.53071928e+00 -8.29922259e-01 3.64875704e-01 -3.38580668e-01
9.71663356e-01 -7.66638458e-01 -8.37664723e-01 4.34263468e-01
2.55872846e-01 9.57102776e-02 1.75953478e-01 -6.54433370e-01
-7.05198944e-01 -7.29154706e-01 -1.50349116e+00 6.52618527e-01
1.36664903e+00 -1.24618828e+00 -1.90311626e-01 1.25832021e-01
4.07097399e-01 -1.02750167e-01 -1.08697462e+00 4.42427814e-01
8.91068459e-01 -5.60905755e-01 9.15996552e-01 -1.86556995e-01
1.86828509e-01 -8.03387687e-02 -1.34356737e-01 -1.03657639e+00
-1.80888534e-01 -3.58635008e-01 -5.47519922e-01 5.89742184e-01
5.36351740e-01 -1.24952888e+00 5.25765896e-01 5.69976747e-01
-8.35749973e-03 -4.38455164e-01 -1.65289164e+00 -1.39234293e+00
-2.16086116e-02 8.07058290e-02 2.09082186e-01 2.68738210e-01
5.70848942e-01 3.45185399e-01 5.56969643e-02 -3.51858824e-01
4.76897240e-01 -1.12645730e-01 -3.94869268e-01 -1.00549674e+00
-2.21885383e-01 -8.34051847e-01 -1.13043952e+00 -9.75366056e-01
1.94026500e-01 -1.06675446e+00 3.74781154e-02 -9.58796918e-01
3.68737102e-01 -3.99890542e-01 -7.05559850e-01 4.23358083e-01
7.22944498e-01 1.17707491e+00 -1.92235768e-01 5.08550763e-01
-7.41110742e-01 3.60400945e-01 8.75154495e-01 -6.87639773e-01
1.53830186e-01 -7.38114953e-01 3.69193517e-02 -3.74554157e-01
6.63302600e-01 -4.98699188e-01 -2.67130047e-01 -9.29940760e-01
2.42808834e-02 2.57237494e-01 5.47563434e-01 -1.94045997e+00
8.86667430e-01 4.54659104e-01 8.92922699e-01 1.49701282e-01
6.37921691e-01 -3.59666616e-01 2.74635255e-01 1.19831991e+00
-3.48330379e-01 7.30747044e-01 6.12997949e-01 4.00836706e-01
8.28176215e-02 1.82781875e-01 6.36392176e-01 2.03889444e-01
-6.42987669e-01 1.94462702e-01 -1.26094556e+00 -2.11178988e-01
5.33539891e-01 -6.29598677e-01 -1.22280133e+00 2.24809065e-01
-4.57449377e-01 -1.20526232e-01 4.22735900e-01 2.61821568e-01
1.30333567e+00 -1.22110224e+00 5.21736220e-02 6.65541649e-01
-2.90682912e-01 -1.19825053e+00 1.67600930e-01 7.07330048e-01
-6.50886655e-01 7.71404266e-01 -1.19324911e+00 -7.99761355e-01
-5.26792824e-01 3.61760139e-01 7.46963739e-01 4.55615848e-01
-6.32930636e-01 6.68230057e-01 -6.42460108e-01 -1.42927215e-01
3.61926146e-02 -8.12453851e-02 3.83156180e-01 -1.01062739e+00
3.07602972e-01 3.16375613e-01 3.96714240e-01 3.52286324e-02
-5.61768234e-01 3.36293548e-01 -8.82560015e-02 -3.69149446e-01
9.25112188e-01 2.66084135e-01 -8.22130620e-01 5.80412984e-01
9.47417498e-01 -8.47628295e-01 -7.28554428e-01 1.39029920e-01
1.20223045e-01 1.73407257e-01 -7.92431980e-02 -9.05457795e-01
-1.33808374e+00 8.95731807e-01 1.15072107e+00 -7.02230483e-02
9.81800795e-01 -4.36544627e-01 1.38997066e+00 1.47936583e+00
9.57856476e-01 -1.26277375e+00 3.96496147e-01 8.05286884e-01
3.53060871e-01 -5.07959008e-01 -6.14210010e-01 3.81867439e-01
7.22106397e-02 1.40456963e+00 1.15505266e+00 -7.42970526e-01
8.29963267e-01 9.79595840e-01 -1.52106658e-01 -4.62064743e-01
-1.16482592e+00 6.55718148e-01 -3.43990892e-01 7.16469109e-01
1.83849439e-01 4.13072914e-01 -1.06504641e-01 3.94648835e-02
-1.64718419e-01 -2.61764228e-02 7.26968110e-01 1.08528471e+00
-6.40163660e-01 -6.45202637e-01 -3.72180641e-02 1.07720149e+00
-2.73309350e-01 -4.29732174e-01 -4.47031140e-01 1.93616316e-01
-7.95566440e-01 8.44264507e-01 6.46713972e-01 -1.12492263e+00
5.62448092e-02 2.12788850e-01 1.07489502e+00 1.23900943e-01
-7.76750267e-01 -7.07317352e-01 -4.43651378e-01 -4.52925563e-01
4.40383971e-01 4.49004710e-01 -1.79350555e+00 -9.29851294e-01
-8.07877854e-02 4.65212092e-02 1.27911270e+00 5.20765543e-01
7.50529170e-01 6.95129216e-01 5.48571885e-01 -4.81025577e-01
-7.51483083e-01 -7.78596878e-01 -9.49021101e-01 -9.46549252e-02
3.32600713e-01 -4.15240765e-01 1.11721054e-01 -5.81148982e-01]
|
[8.218255996704102, 2.514755964279175]
|
ef86ea61-741d-4311-9c76-dfcbb6d028be
|
period-vits-variational-inference-with
|
2210.15964
| null |
https://arxiv.org/abs/2210.15964v2
|
https://arxiv.org/pdf/2210.15964v2.pdf
|
Period VITS: Variational Inference with Explicit Pitch Modeling for End-to-end Emotional Speech Synthesis
|
Several fully end-to-end text-to-speech (TTS) models have been proposed that have shown better performance compared to cascade models (i.e., training acoustic and vocoder models separately). However, they often generate unstable pitch contour with audible artifacts when the dataset contains emotional attributes, i.e., large diversity of pronunciation and prosody. To address this problem, we propose Period VITS, a novel end-to-end TTS model that incorporates an explicit periodicity generator. In the proposed method, we introduce a frame pitch predictor that predicts prosodic features, such as pitch and voicing flags, from the input text. From these features, the proposed periodicity generator produces a sample-level sinusoidal source that enables the waveform decoder to accurately reproduce the pitch. Finally, the entire model is jointly optimized in an end-to-end manner with variational inference and adversarial objectives. As a result, the decoder becomes capable of generating more stable, expressive, and natural output waveforms. The experimental results showed that the proposed model significantly outperforms baseline models in terms of naturalness, with improved pitch stability in the generated samples.
|
['Kentaro Tachibana', 'Jae-Min Kim', 'Ryo Terashima', 'Eunwoo Song', 'Ryuichi Yamamoto', 'Yuma Shirahata']
|
2022-10-28
| null | null | null | null |
['emotional-speech-synthesis']
|
['speech']
|
[-1.79263443e-01 2.12460116e-01 1.57724231e-01 -2.28325188e-01
-1.10340858e+00 -5.67108393e-01 3.15371484e-01 -2.74587393e-01
9.68415141e-02 7.13921428e-01 4.27904755e-01 1.11239962e-01
6.20184064e-01 -4.84809011e-01 -6.88172162e-01 -7.89167404e-01
3.89074981e-01 -4.68968637e-02 -1.96913220e-02 -2.05246657e-01
-1.24635279e-01 -7.34184310e-02 -1.49012148e+00 8.30777511e-02
9.56205666e-01 1.12389553e+00 3.61744165e-01 8.13182950e-01
1.73409984e-01 5.01360238e-01 -9.38624203e-01 -4.24119025e-01
-3.52251977e-02 -7.52953351e-01 5.11574820e-02 -9.31017026e-02
-4.83541749e-03 -2.39263520e-01 -2.81830400e-01 1.12761116e+00
9.36600864e-01 2.89076209e-01 6.31529331e-01 -9.12354887e-01
-6.04671538e-01 7.33911216e-01 -2.22409099e-01 -2.10926518e-01
2.46138319e-01 3.90310496e-01 1.13767159e+00 -1.14856684e+00
3.58872950e-01 1.38119090e+00 5.28772712e-01 5.79811394e-01
-1.26316369e+00 -9.04486060e-01 -1.12967871e-01 8.74552131e-02
-1.30798018e+00 -7.24094987e-01 1.37828088e+00 -3.56763482e-01
7.05427170e-01 2.14175344e-01 5.59001148e-01 1.50975811e+00
1.78225324e-01 8.10870945e-01 7.45856464e-01 -3.85571063e-01
2.42566928e-01 2.27563053e-01 -4.94394422e-01 2.28027344e-01
-5.63870490e-01 3.88728946e-01 -5.66117287e-01 -5.74187301e-02
5.10641277e-01 -6.42348945e-01 -3.41666818e-01 3.03165168e-01
-9.19003129e-01 7.04034686e-01 8.27984959e-02 1.16199896e-01
-5.46012700e-01 1.07517384e-01 5.34303486e-01 -2.27850862e-02
4.30976123e-01 2.54575104e-01 -1.24374077e-01 -2.67605990e-01
-1.02084398e+00 4.38891947e-01 6.36012375e-01 7.58069754e-01
2.86906391e-01 8.96218479e-01 -5.18286943e-01 1.08046126e+00
4.80596244e-01 8.28359663e-01 6.88538492e-01 -8.96182835e-01
5.03472328e-01 -2.01170549e-01 2.40150705e-01 -9.30075645e-01
-1.83934286e-01 -5.19118965e-01 -8.01654041e-01 6.92647845e-02
-1.91752650e-02 -5.75058579e-01 -6.98017836e-01 2.11376834e+00
3.47040176e-01 3.21444154e-01 5.56472838e-01 9.93829072e-01
9.20344591e-01 1.23487282e+00 1.23680837e-01 -5.51777422e-01
1.26113486e+00 -9.60542500e-01 -1.55181575e+00 -4.16075252e-02
-1.85577095e-01 -9.87643540e-01 1.34431100e+00 4.09313709e-01
-1.44614267e+00 -1.02683699e+00 -1.11545300e+00 3.11798789e-02
1.26042128e-01 3.10479939e-01 -2.55610794e-01 5.25460958e-01
-7.67938375e-01 5.60493410e-01 -5.89740098e-01 3.48062664e-01
-2.00209439e-01 -1.96136720e-02 3.57190907e-01 7.30453014e-01
-1.57629454e+00 4.67918307e-01 4.86157417e-01 2.75893025e-02
-8.98478806e-01 -7.03830779e-01 -9.84360993e-01 2.43770078e-01
1.35977566e-01 -4.32303965e-01 1.68866897e+00 -9.81300533e-01
-2.30537152e+00 6.89940602e-02 -3.78391087e-01 -4.17884201e-01
3.61674458e-01 -3.04133832e-01 -7.67432332e-01 1.88823596e-01
-1.27831563e-01 7.49669373e-01 1.22383785e+00 -1.30370855e+00
-5.37956536e-01 2.29436725e-01 -6.36830866e-01 4.18465286e-01
-3.00217658e-01 -8.76024961e-02 -2.73463309e-01 -1.07170534e+00
-3.32311571e-01 -7.14848638e-01 1.16109643e-02 -3.47477049e-01
-6.71247721e-01 -2.03410208e-01 7.70399272e-01 -9.72386420e-01
1.42874432e+00 -2.55148292e+00 1.64427698e-01 -9.35651436e-02
-2.60567099e-01 3.83993208e-01 1.39020428e-01 4.27134603e-01
8.08753222e-02 9.13240686e-02 -1.28090113e-01 -7.15141237e-01
2.57796109e-01 -8.77698362e-02 -7.19385684e-01 -3.09877545e-02
5.03265798e-01 7.21778512e-01 -8.61111879e-01 -5.61233580e-01
3.43269050e-01 7.24604845e-01 -7.15519965e-01 6.16989613e-01
-5.32473147e-01 7.92225957e-01 -1.10221155e-01 3.61543000e-01
3.78885061e-01 4.31900769e-01 -1.84773266e-01 -1.39766872e-01
-2.20297337e-01 3.77822399e-01 -1.07747877e+00 1.42919433e+00
-6.21266186e-01 5.34141719e-01 1.05558157e-01 -4.24763501e-01
1.29789424e+00 9.42888796e-01 1.28188878e-01 -4.71231103e-01
2.73360163e-01 3.62409562e-01 1.24084994e-01 -4.20167714e-01
5.43353438e-01 -3.73848110e-01 -1.17474936e-01 -1.22948959e-01
-3.11948941e-03 -5.49997807e-01 -1.82676271e-01 -3.41391921e-01
4.26637292e-01 6.96793273e-02 2.26814225e-02 1.70914188e-01
6.63979053e-01 -4.63479638e-01 8.33791375e-01 8.94563645e-02
-9.02552828e-02 8.94788146e-01 3.31212580e-01 2.47785285e-01
-1.20645833e+00 -1.25446832e+00 -1.86852142e-02 7.13970780e-01
1.22489464e-02 -2.48935193e-01 -1.03522873e+00 6.64024502e-02
-2.06692457e-01 1.14176428e+00 -2.17777684e-01 -2.44166791e-01
-4.99648035e-01 -2.43916705e-01 8.13855767e-01 3.71623367e-01
4.03550357e-01 -1.45754004e+00 -1.35694742e-01 7.54189610e-01
-4.46138620e-01 -1.15938950e+00 -9.93667901e-01 -1.39047593e-01
-5.37122548e-01 -1.75063133e-01 -9.18022335e-01 -8.47913623e-01
1.52735993e-01 -4.11532670e-01 6.44302130e-01 -5.87679327e-01
1.23300269e-01 -1.49616286e-01 -4.13944066e-01 -5.77636242e-01
-7.88486123e-01 -2.50014603e-01 2.74370760e-01 4.15816665e-01
-2.35274687e-01 -7.98386216e-01 -6.38564467e-01 2.07493007e-01
-7.78525889e-01 7.42061958e-02 3.94947916e-01 9.34573829e-01
9.24837470e-01 1.21061303e-01 1.21875238e+00 -4.33663368e-01
1.02658212e+00 -4.10489291e-01 -4.84081388e-01 -1.61353484e-01
-3.08397144e-01 -1.48121625e-01 1.37093580e+00 -8.54726136e-01
-1.45441794e+00 -9.03282315e-02 -6.14829659e-01 -9.43248868e-01
-5.27950062e-04 4.00834829e-01 -5.72977901e-01 5.50820053e-01
5.16064882e-01 4.38970506e-01 -1.67784672e-02 -3.97287637e-01
4.87749368e-01 9.98596430e-01 9.74458992e-01 -3.85262281e-01
8.41683507e-01 -2.02777579e-01 -4.57578629e-01 -1.00106347e+00
-5.41535139e-01 -4.94770519e-02 -5.33516668e-02 -2.22385332e-01
8.63625765e-01 -1.12127066e+00 -6.76326692e-01 7.30339229e-01
-1.37291467e+00 -2.97115624e-01 -2.84476370e-01 7.08617449e-01
-8.74303401e-01 2.81347632e-01 -8.55202258e-01 -1.25926542e+00
-7.34871864e-01 -1.17672038e+00 1.02700567e+00 5.45636237e-01
-4.06571746e-01 -6.64633393e-01 1.99497998e-01 1.34998232e-01
3.44385266e-01 6.07063413e-01 7.78275669e-01 -2.51432151e-01
-1.42534506e-02 1.31276011e-01 4.53577727e-01 6.97911382e-01
1.70945361e-01 3.01196158e-01 -1.30093288e+00 -4.33414206e-02
1.94671988e-01 -2.94322014e-01 3.87231886e-01 5.22528291e-01
1.00197482e+00 -7.05343306e-01 1.74150005e-01 5.93962073e-01
1.14648139e+00 6.83603346e-01 6.74255669e-01 -3.80469054e-01
4.51752901e-01 5.02092540e-01 8.00148964e-01 5.93204021e-01
5.51536493e-02 5.88903308e-01 1.84491500e-01 -1.54903159e-01
-3.68077517e-01 -6.30010128e-01 7.68694162e-01 1.51895130e+00
3.21153104e-01 -5.20158529e-01 -3.52856010e-01 5.40282071e-01
-1.54006374e+00 -9.76257503e-01 1.11311264e-02 1.92859852e+00
1.26455081e+00 2.62652695e-01 1.89918816e-01 2.26770788e-01
9.35053110e-01 3.01440239e-01 -8.88075829e-01 -7.17746079e-01
-1.22487336e-01 2.60257632e-01 -2.60632150e-02 5.11519372e-01
-7.29114056e-01 9.83187795e-01 5.45379686e+00 1.04997599e+00
-1.27031255e+00 -1.11431606e-01 4.17533964e-01 -2.07748696e-01
-5.08542717e-01 -4.05782223e-01 -6.88667357e-01 8.26934278e-01
1.13329220e+00 -4.47719306e-01 5.90683281e-01 8.05947125e-01
7.65886664e-01 4.24546093e-01 -8.85133386e-01 8.32358003e-01
-1.65488064e-01 -8.88565660e-01 -4.27409485e-02 -2.72147179e-01
6.82924330e-01 -4.81623560e-01 4.86688673e-01 4.37937260e-01
-9.10261944e-02 -9.70794082e-01 1.32306325e+00 4.03124690e-01
1.14952064e+00 -1.02417588e+00 4.64503974e-01 4.55419481e-01
-1.31907928e+00 1.95942577e-02 -1.12872496e-01 1.26623452e-01
5.95814824e-01 4.09732640e-01 -1.03643978e+00 4.91772771e-01
1.95441067e-01 3.07179034e-01 1.69819251e-01 8.50732148e-01
-4.41255420e-01 1.03004432e+00 -9.82766971e-02 -1.92809671e-01
1.43061757e-01 -8.72044489e-02 8.26776266e-01 1.29683220e+00
5.06719589e-01 2.10509151e-01 7.26819634e-02 1.09676707e+00
-1.23929188e-01 3.17586750e-01 -4.11827713e-01 -2.50451833e-01
8.85998249e-01 1.06611383e+00 -1.12487972e-01 -7.01096058e-02
-1.41562939e-01 7.99657822e-01 -7.26083368e-02 5.45255482e-01
-1.19809496e+00 -7.01387584e-01 6.83515489e-01 -2.51681060e-01
4.05501068e-01 -3.18545476e-02 -1.22785576e-01 -8.46033514e-01
1.57271311e-01 -8.58717144e-01 -1.78473145e-01 -8.59768331e-01
-1.11803162e+00 7.78851330e-01 -3.82301182e-01 -1.27293801e+00
-6.42530620e-01 -1.25737265e-01 -1.02660882e+00 1.20837843e+00
-1.48267174e+00 -8.33705544e-01 6.98881745e-02 4.36954707e-01
8.94844592e-01 -8.88115317e-02 7.63257444e-01 2.48762354e-01
-5.89071929e-01 8.53530645e-01 2.07364380e-01 7.37462342e-02
7.82225192e-01 -1.25023150e+00 5.72794676e-01 7.87807226e-01
-1.27529815e-01 1.91423848e-01 1.05985785e+00 -4.92717445e-01
-1.06076527e+00 -1.35407078e+00 6.42142892e-01 5.41578978e-02
4.79917973e-01 -4.96198267e-01 -1.02225220e+00 1.58791795e-01
4.46165144e-01 -3.27807993e-01 5.84177732e-01 -4.60361511e-01
1.56800404e-01 -1.65426299e-01 -9.95042026e-01 7.77535796e-01
4.61265415e-01 -5.20314693e-01 -6.43718958e-01 -1.65991530e-01
1.27417767e+00 -7.00119615e-01 -7.98129439e-01 3.31352293e-01
4.72127199e-01 -7.73295343e-01 6.77270710e-01 -5.27850837e-02
4.44486201e-01 -4.12492782e-01 -1.33926958e-01 -1.54727232e+00
-1.50641516e-01 -1.21842194e+00 -2.98695743e-01 1.76217675e+00
5.07188857e-01 -3.50078553e-01 2.21426025e-01 2.03063905e-01
-4.88430202e-01 -7.61949599e-01 -8.70129824e-01 -8.69596660e-01
4.79594208e-02 -3.53510886e-01 5.44699669e-01 3.98830235e-01
4.84590009e-02 5.50791264e-01 -8.18149924e-01 3.81456673e-01
3.92337948e-01 -5.03628850e-02 6.45739198e-01 -8.03344011e-01
-4.83998001e-01 -4.99488801e-01 2.04779863e-01 -1.06795835e+00
2.13327900e-01 -6.22954667e-01 6.47347569e-01 -1.05558181e+00
-4.33383107e-01 -1.20226577e-01 -2.40585327e-01 -1.38701703e-02
-5.76797009e-01 -1.37983471e-01 4.36652333e-01 -3.21641117e-02
1.72913484e-02 1.19969094e+00 1.44553387e+00 4.16277302e-03
-5.19193113e-01 2.25052744e-01 -4.40059990e-01 6.28623128e-01
9.86719549e-01 -4.27091748e-01 -5.68121493e-01 -6.15960360e-02
-3.66227865e-01 7.25871503e-01 7.55203441e-02 -1.00762010e+00
3.01714558e-02 1.93465147e-02 2.42214367e-01 -6.69074655e-01
8.73838007e-01 -4.20185924e-01 2.01546460e-01 3.00673276e-01
-5.15433669e-01 -3.07953179e-01 3.29870790e-01 5.65933704e-01
-5.42507708e-01 -7.47396052e-02 9.41147685e-01 2.64435679e-01
-2.99001858e-02 2.10209265e-01 -3.86006713e-01 2.45062515e-01
8.12294304e-01 3.31525914e-02 7.66253024e-02 -6.27268970e-01
-6.07130229e-01 2.03067988e-01 2.80163307e-02 5.32142103e-01
6.04620337e-01 -1.56875885e+00 -1.03133190e+00 2.00308636e-01
-2.94703245e-01 1.35730430e-01 4.78237629e-01 3.41206402e-01
6.59759864e-02 3.29756707e-01 1.32165298e-01 -4.67050076e-01
-9.55427706e-01 4.67902184e-01 2.45805159e-01 -3.18951420e-02
-5.79515040e-01 5.75850964e-01 3.31212670e-01 -1.35330230e-01
5.80607295e-01 -4.97990698e-01 -2.68154502e-01 1.33340865e-01
3.71022850e-01 5.71573190e-02 -3.02813888e-01 -6.17955387e-01
-9.75794718e-02 4.42225933e-01 4.10923302e-01 -7.39417255e-01
9.40886140e-01 -1.83232740e-01 5.16774535e-01 8.10887158e-01
1.16249788e+00 4.68297005e-01 -1.68544018e+00 8.28086585e-03
-4.68375474e-01 -1.64016664e-01 -1.01817353e-02 -7.43332148e-01
-9.12259161e-01 1.06069326e+00 3.00450653e-01 9.32991281e-02
1.24762106e+00 -3.75068635e-01 1.51391757e+00 -1.21623702e-01
-4.15437408e-02 -1.28973544e+00 2.43797824e-01 5.26463985e-01
1.05398524e+00 -8.48958552e-01 -7.54400253e-01 -1.97133392e-01
-1.11279297e+00 1.03830981e+00 5.24118483e-01 -5.57578430e-02
4.08512205e-01 2.41244733e-01 2.81507254e-01 5.04204035e-01
-8.83071840e-01 1.08594716e-01 2.68909484e-01 5.56893051e-01
2.73205638e-01 2.35124618e-01 -2.37264663e-01 1.13308609e+00
-8.64561677e-01 -2.45093942e-01 3.98423493e-01 1.93024591e-01
-4.60566938e-01 -8.60549450e-01 -5.22216022e-01 -1.62839174e-01
-5.75799167e-01 -1.93269357e-01 -1.71754032e-01 3.75628233e-01
6.12711580e-03 1.19280255e+00 -2.33159587e-02 -5.48452556e-01
6.20697021e-01 3.62899065e-01 -3.72894220e-02 -3.57000828e-01
-7.06377983e-01 7.37680852e-01 9.31111425e-02 -1.43091172e-01
2.32633308e-01 -5.70601046e-01 -1.53612709e+00 9.76798981e-02
-4.52675134e-01 3.55396152e-01 5.68639994e-01 5.61736822e-01
2.92816013e-01 9.80857193e-01 1.29485881e+00 -8.49739492e-01
-7.89641559e-01 -1.17418993e+00 -6.51444852e-01 3.97397876e-01
5.89978099e-01 -3.04244280e-01 -5.03366947e-01 3.13289851e-01]
|
[15.290383338928223, 6.273355007171631]
|
3ff40401-d3e6-4263-8ed8-e659bf189697
|
privacy-preserving-community-detection-for
|
2306.15709
| null |
https://arxiv.org/abs/2306.15709v1
|
https://arxiv.org/pdf/2306.15709v1.pdf
|
Privacy-Preserving Community Detection for Locally Distributed Multiple Networks
|
Modern multi-layer networks are commonly stored and analyzed in a local and distributed fashion because of the privacy, ownership, and communication costs. The literature on the model-based statistical methods for community detection based on these data is still limited. This paper proposes a new method for consensus community detection and estimation in a multi-layer stochastic block model using locally stored and computed network data with privacy protection. A novel algorithm named privacy-preserving Distributed Spectral Clustering (ppDSC) is developed. To preserve the edges' privacy, we adopt the randomized response (RR) mechanism to perturb the network edges, which satisfies the strong notion of differential privacy. The ppDSC algorithm is performed on the squared RR-perturbed adjacency matrices to prevent possible cancellation of communities among different layers. To remove the bias incurred by RR and the squared network matrices, we develop a two-step bias-adjustment procedure. Then we perform eigen-decomposition on the debiased matrices, aggregation of the local eigenvectors using an orthogonal Procrustes transformation, and k-means clustering. We provide theoretical analysis on the statistical errors of ppDSC in terms of eigen-vector estimation. In addition, the blessings and curses of network heterogeneity are well-explained by our bounds.
|
['Shujie Ma', 'Xiangyu Chang', 'Xiang Li', 'Xiao Guo']
|
2023-06-27
| null | null | null | null |
['stochastic-block-model', 'community-detection', 'clustering']
|
['graphs', 'graphs', 'methodology']
|
[ 2.10371405e-01 -4.14662719e-01 4.86166887e-02 -1.33116126e-01
-6.37993097e-01 -8.06055725e-01 1.35273412e-01 7.37070367e-02
-2.90158272e-01 6.39137805e-01 2.27254763e-01 -2.12917373e-01
-5.39415240e-01 -6.53273880e-01 -4.70121115e-01 -1.26856518e+00
-5.39748013e-01 8.11657533e-02 1.45667031e-01 2.97864854e-01
3.05761278e-01 6.04242444e-01 -8.73763621e-01 3.25332463e-01
7.75345504e-01 7.56792784e-01 -4.99164551e-01 6.29834890e-01
1.35074973e-01 6.46108329e-01 -3.50436509e-01 -5.70831716e-01
5.92707634e-01 -3.82766664e-01 -3.91819090e-01 1.53005570e-01
-1.18282370e-01 -1.42013788e-01 -3.89066905e-01 1.36380696e+00
5.33144057e-01 -3.53015006e-01 6.82330847e-01 -1.72805643e+00
-4.06123668e-01 8.09112489e-01 -1.17027009e+00 -2.10642725e-01
1.68065771e-01 -2.44087800e-01 7.93864727e-01 -6.23901904e-01
6.75442457e-01 1.21704221e+00 8.91100645e-01 3.97594094e-01
-1.73902142e+00 -1.03802097e+00 -1.32938728e-01 7.62999654e-02
-2.15861535e+00 -5.33270240e-01 9.85455036e-01 -5.62468052e-01
8.62001106e-02 8.00674379e-01 3.77204061e-01 8.01218510e-01
1.86132401e-01 4.10662770e-01 1.19410419e+00 -1.75261900e-01
4.62946951e-01 2.20309123e-01 9.26129869e-04 4.16128993e-01
9.71170962e-01 -8.81794915e-02 -5.37332714e-01 -1.08388317e+00
6.44243717e-01 1.66953519e-01 -3.81945640e-01 -1.07550466e+00
-9.47280407e-01 6.99092984e-01 9.68539491e-02 1.28606007e-01
-3.92080069e-01 2.11331159e-01 4.59109515e-01 4.52693373e-01
6.39828861e-01 -3.65344763e-01 -2.27658838e-01 5.68497956e-01
-1.13737392e+00 -1.12104975e-01 1.07079756e+00 9.54516888e-01
6.80892408e-01 -4.63140488e-01 3.30892508e-03 3.36827576e-01
6.15951240e-01 8.20677757e-01 -1.98553145e-01 -1.06969130e+00
4.69322801e-01 4.95572686e-01 7.96619356e-02 -1.81860280e+00
-1.27856508e-01 -3.75425547e-01 -1.82855690e+00 -5.69943525e-02
2.19554946e-01 -4.39023554e-01 6.14080243e-02 1.74019158e+00
4.95114982e-01 1.19517565e-01 2.25274153e-02 6.47086620e-01
1.52036220e-01 4.45510149e-01 -3.85303706e-01 -7.48662114e-01
8.06321084e-01 -2.51066387e-01 -7.45934248e-01 3.76139045e-01
4.54105824e-01 -3.23342800e-01 1.25882000e-01 1.76987737e-01
-7.62068510e-01 4.10072714e-01 -8.56169879e-01 4.52494711e-01
7.28464499e-02 -2.44752094e-01 3.51813257e-01 1.15090895e+00
-1.38352382e+00 3.33656579e-01 -8.59077990e-01 -2.85394609e-01
4.48064625e-01 4.48620558e-01 -6.16062462e-01 -1.19622275e-01
-8.59724820e-01 1.56011671e-01 -1.44991249e-01 2.93556005e-01
-4.66306776e-01 -4.90245104e-01 -4.71428543e-01 6.72407672e-02
2.94318736e-01 -4.67851728e-01 1.45944878e-01 -8.45177531e-01
-1.03529871e+00 7.46802211e-01 -1.57458916e-01 -5.05062103e-01
7.87978947e-01 5.11583328e-01 -2.98544824e-01 2.35793203e-01
2.97253486e-02 -2.89226711e-01 8.85079205e-01 -1.51550543e+00
-1.38894066e-01 -5.97247541e-01 -8.04237783e-01 -1.72533482e-01
-4.95204300e-01 2.49435022e-01 -2.95745134e-01 -6.27871692e-01
6.44271970e-01 -9.45372701e-01 -5.52684784e-01 1.43555552e-01
-6.41034365e-01 3.98888052e-01 5.90965986e-01 -7.46851683e-01
1.65426815e+00 -2.30594659e+00 1.03184201e-01 1.22823644e+00
6.36968613e-01 -2.67952472e-01 -1.58956707e-01 7.87162483e-01
-2.73882691e-02 3.84112358e-01 -4.56127584e-01 -4.37562019e-01
-1.19412191e-01 -2.61383355e-01 -5.85376844e-02 1.21632302e+00
-5.54100990e-01 2.64988452e-01 -5.09546578e-01 -5.30654252e-01
-3.01246017e-01 2.62243420e-01 -7.05305994e-01 -2.01487377e-01
5.41470110e-01 2.40996197e-01 -4.38185722e-01 5.10248899e-01
1.37165368e+00 -4.41408426e-01 9.46859539e-01 -1.52088225e-01
-5.14939763e-02 -3.89464170e-01 -1.82918882e+00 1.10798383e+00
2.85298437e-01 3.21978569e-01 1.13694143e+00 -7.84979641e-01
7.73997128e-01 5.29511511e-01 7.87864506e-01 1.77524656e-01
9.91247892e-02 2.26830736e-01 -1.13959461e-01 3.03068664e-02
7.09884316e-02 1.63440958e-01 -1.53631538e-01 8.82191658e-01
-5.43377697e-01 4.69358861e-01 -2.65631646e-01 6.89057708e-01
1.45401692e+00 -8.66433203e-01 2.23713174e-01 -6.26837492e-01
5.92486262e-01 -4.21810061e-01 8.63655388e-01 8.29786181e-01
-3.64695877e-01 4.33731943e-01 6.83772743e-01 9.94287953e-02
-9.27231491e-01 -8.94754827e-01 9.88894254e-02 5.58091164e-01
2.20607221e-01 -5.98371506e-01 -9.07308698e-01 -3.56242001e-01
4.28233057e-01 -2.79539414e-02 -4.20922577e-01 -3.47951241e-02
4.68309335e-02 -1.02668214e+00 6.73192859e-01 -2.46126438e-03
4.43834007e-01 -3.05864483e-01 1.23390950e-01 8.76239017e-02
-2.86706775e-01 -6.87102914e-01 -7.60203421e-01 -2.77551174e-01
-8.20678353e-01 -1.19935763e+00 -3.42752337e-01 -5.00194192e-01
9.45850611e-01 6.41180158e-01 4.71121907e-01 1.56733826e-01
-3.92215922e-02 5.92760444e-01 -9.54248570e-03 1.21049017e-01
-3.78416985e-01 -2.00583905e-01 4.54861522e-01 8.39660347e-01
3.37468088e-01 -9.45532918e-01 -7.10565448e-01 3.03421378e-01
-7.18390405e-01 -1.91773251e-01 4.76970583e-01 5.04840016e-01
3.53449404e-01 4.28787768e-01 1.95811629e-01 -8.87914419e-01
9.15150583e-01 -6.91439450e-01 -8.36842299e-01 3.63397121e-01
-7.62093961e-01 -2.68520005e-02 2.20428407e-01 -2.49409437e-01
-8.31952512e-01 3.10699999e-01 6.65416241e-01 -4.15007442e-01
3.53408635e-01 4.40354973e-01 -3.54821175e-01 -5.44187009e-01
4.00395781e-01 2.16293141e-01 3.46716076e-01 -5.15521288e-01
3.48857641e-01 1.06481493e+00 1.63253874e-01 -5.18851936e-01
1.05594826e+00 8.82311463e-01 2.02753201e-01 -7.87367165e-01
6.65672943e-02 -6.51085973e-01 -4.80394155e-01 -2.55780369e-01
4.26496655e-01 -1.10120380e+00 -1.13462842e+00 7.71514356e-01
-1.10501993e+00 8.13919157e-02 2.68996596e-01 2.80078858e-01
-6.40318692e-02 1.01431704e+00 -7.45375991e-01 -1.32153785e+00
-6.22950435e-01 -5.92662513e-01 3.17668766e-01 -3.17345560e-01
-5.47576323e-02 -7.55815029e-01 9.75149050e-02 1.11915253e-01
4.35356349e-01 3.25709760e-01 5.43520868e-01 -5.68611801e-01
-7.82313645e-01 -4.78711605e-01 -3.68896365e-01 3.42045695e-01
-2.43807603e-02 2.35988021e-01 -4.89241332e-01 -8.66294563e-01
1.25540018e-01 3.33238512e-01 4.92836565e-01 4.57943648e-01
1.17238498e+00 -6.92990601e-01 -5.34171283e-01 6.48319304e-01
1.45878243e+00 -2.63087094e-01 4.84537810e-01 -9.56508145e-03
6.41800821e-01 6.26979530e-01 -5.78225404e-02 1.08238864e+00
3.43337923e-01 9.48164612e-02 2.96527684e-01 1.38771713e-01
6.06440306e-01 -2.76241779e-01 3.50839198e-01 1.25184178e+00
1.74535766e-01 -3.43104959e-01 -8.12392592e-01 5.75473726e-01
-2.03036571e+00 -1.04059494e+00 -6.56351745e-01 2.38683748e+00
8.66189063e-01 -2.32530892e-01 5.51987737e-02 8.90433490e-02
1.31687367e+00 9.76893976e-02 -3.82877171e-01 1.25291586e-01
-5.21556139e-01 -3.60085279e-01 1.39363074e+00 4.42923069e-01
-9.20953631e-01 3.90296042e-01 6.23615408e+00 8.15593600e-01
-5.69776773e-01 1.76001832e-01 5.73018014e-01 5.30846454e-02
-4.36185390e-01 2.55810857e-01 -2.98817486e-01 5.51421881e-01
7.94360757e-01 -3.78228873e-01 7.80645490e-01 5.11497855e-01
3.78528774e-01 9.02264845e-03 -8.54242563e-01 1.20160592e+00
-1.00684330e-01 -1.31707907e+00 -1.81662843e-01 6.04972303e-01
1.02328825e+00 -1.89714402e-01 -1.41342044e-01 -5.57357371e-01
6.92668080e-01 -5.71706891e-01 3.48454893e-01 7.31513798e-01
8.76868188e-01 -6.87139750e-01 5.24360597e-01 3.98734421e-01
-1.16396475e+00 -1.26152471e-01 -4.47562873e-01 2.83198301e-02
-1.34045780e-01 9.45290267e-01 -1.49207309e-01 5.40019870e-01
6.89258158e-01 6.13711357e-01 -3.13758343e-01 9.56992447e-01
2.12812200e-01 6.91220820e-01 -8.00288439e-01 -2.53269989e-02
-3.78777862e-01 -7.68059492e-01 8.10616195e-01 9.46454167e-01
3.66614580e-01 1.73329175e-01 -1.12947315e-01 7.31731892e-01
-3.56878757e-01 2.52102643e-01 -4.35430318e-01 9.18060690e-02
1.21813095e+00 1.04516625e+00 -5.05590320e-01 -8.63066688e-02
-3.14266920e-01 1.06394184e+00 -1.00225722e-02 5.11121035e-01
-2.25079104e-01 -3.47378552e-01 7.87212968e-01 2.85213739e-01
2.86358953e-01 -2.05187738e-01 -5.60794830e-01 -1.24997520e+00
1.10816970e-01 -1.00940454e+00 4.57199961e-01 -2.78018236e-01
-1.73304796e+00 -9.36987475e-02 -2.51221120e-01 -9.89414155e-01
3.12478364e-01 5.88889141e-03 -5.22436619e-01 7.72049725e-01
-7.53964305e-01 -6.55452967e-01 5.50220795e-02 9.98734653e-01
-7.04670310e-01 -2.08355233e-01 6.58563495e-01 3.99749368e-01
-7.95495212e-01 7.94624865e-01 1.06032717e+00 3.18321884e-01
8.23605478e-01 -7.39959598e-01 -3.59291546e-02 1.14782643e+00
-2.50888050e-01 9.00246084e-01 6.16112769e-01 -8.42045069e-01
-1.48110330e+00 -1.01052570e+00 9.12896991e-01 -4.54000533e-02
7.64708996e-01 -8.10209572e-01 -8.60412002e-01 6.42788708e-01
-1.03930943e-01 1.43981814e-01 9.26491559e-01 -8.72828290e-02
-4.97499198e-01 -6.10182047e-01 -1.58907020e+00 5.58645725e-01
9.97418463e-01 -7.92144418e-01 2.09653243e-01 2.70094901e-01
3.39574665e-01 3.59325469e-01 -8.91421020e-01 1.02868065e-01
5.79285502e-01 -8.91014695e-01 8.10407639e-01 -3.01235825e-01
-3.30583692e-01 -4.83202696e-01 -3.26017767e-01 -8.66720498e-01
-5.98549604e-01 -1.23247409e+00 -4.48304303e-02 1.52905202e+00
2.22925141e-01 -7.71711767e-01 1.08205628e+00 7.61957169e-01
9.44442332e-01 7.96732232e-02 -1.10039163e+00 -5.57054341e-01
-1.61811844e-01 -2.76135169e-02 5.51587582e-01 1.21737766e+00
1.64646864e-01 -4.94734794e-02 -7.46898413e-01 5.41649818e-01
1.44851410e+00 -1.14830367e-01 7.94861913e-01 -1.45401990e+00
8.61879904e-03 -1.26412928e-01 -2.72999018e-01 -5.16523600e-01
2.94858627e-02 -8.18786979e-01 -2.34350294e-01 -9.67258215e-01
6.34115040e-01 -4.28766608e-01 -2.82675624e-01 1.82442870e-02
1.75892398e-01 -1.95592567e-02 8.67755339e-02 6.12695813e-01
-7.33293355e-01 2.75062710e-01 5.25290072e-01 6.86819181e-02
-1.30797729e-01 2.00040460e-01 -7.18511224e-01 3.24522614e-01
8.07564318e-01 -9.12510335e-01 -2.29519308e-01 8.38904679e-02
3.36186230e-01 3.46355528e-01 4.17433381e-01 -7.44267344e-01
8.27778578e-01 -1.18006617e-01 7.54444720e-03 -7.27113605e-01
-1.10010266e-01 -1.30278814e+00 8.35713387e-01 7.30466664e-01
-2.80792475e-01 -1.77352466e-02 -3.74408454e-01 1.22574544e+00
1.23626217e-01 2.50444949e-01 7.57952332e-01 1.62879631e-01
1.50673151e-01 3.67913306e-01 -7.75475621e-01 -1.33090377e-01
1.03726447e+00 -1.42505959e-01 -1.49918035e-01 -7.01031268e-01
-5.54726899e-01 4.86119717e-01 7.24974573e-01 -2.80324638e-01
4.68249500e-01 -1.30471790e+00 -1.00273740e+00 3.97503138e-01
-2.27578208e-01 -3.83634597e-01 4.91957158e-01 1.07688332e+00
-4.55806464e-01 9.00157094e-02 4.63237055e-02 -3.83286983e-01
-1.63956141e+00 6.67359352e-01 2.06697404e-01 -1.65359616e-01
-1.47555977e-01 7.19495714e-01 -8.62355456e-02 -3.95291805e-01
4.19048518e-01 3.55028182e-01 4.89976257e-01 6.14900514e-02
3.12155396e-01 7.58494079e-01 -1.44472644e-01 -5.66786826e-01
-7.00749755e-01 3.29596788e-01 5.95846847e-02 -2.04936504e-01
1.26689363e+00 -7.38650024e-01 -9.82727826e-01 4.99188788e-02
1.36411536e+00 4.89300966e-01 -9.62977290e-01 -5.30766129e-01
1.74631700e-01 -4.40487713e-01 -1.24140918e-01 -2.72974819e-01
-1.24030006e+00 3.39463234e-01 4.42504704e-01 3.80721718e-01
9.97762978e-01 -2.88694829e-01 2.17963383e-01 1.23707719e-01
5.17970622e-01 -9.68117177e-01 -5.55310428e-01 1.02185331e-01
6.04167223e-01 -8.00421178e-01 2.19944552e-01 -6.08290493e-01
-2.89277047e-01 6.16376281e-01 -8.65431223e-03 -1.93708375e-01
1.13708389e+00 3.28552186e-01 -2.52685040e-01 -4.60765697e-02
-7.36719549e-01 4.77824986e-01 -2.37742543e-01 6.73692822e-01
-1.77136585e-02 2.72499412e-01 -6.33589268e-01 8.40920568e-01
3.37271959e-01 -3.16844016e-01 5.61060250e-01 7.80896664e-01
-1.77289188e-01 -1.09655666e+00 -7.08301485e-01 3.87879640e-01
-4.62981641e-01 -1.17573708e-01 -8.97912681e-01 2.66113997e-01
-3.34592283e-01 1.11373150e+00 -1.27057135e-01 -5.25544107e-01
-2.92091072e-02 -1.40242428e-01 -1.05123304e-01 -1.78525984e-01
-4.29801434e-01 1.43821627e-01 -8.94542560e-02 -5.07162511e-01
-5.42255640e-01 -9.64969456e-01 -6.63400948e-01 -1.15407538e+00
-4.17328179e-01 4.71145391e-01 4.81676698e-01 3.12667638e-01
8.17542672e-01 -2.15109870e-01 1.18946719e+00 -3.60664994e-01
-8.35917413e-01 -5.46193838e-01 -1.22227085e+00 3.27424377e-01
3.10090154e-01 -5.43980449e-02 -9.92988288e-01 -2.12634534e-01]
|
[6.860136985778809, 5.296004295349121]
|
c6965e70-1a38-43c5-965b-b249ecb145ef
|
few-shot-action-recognition-with-compromised
|
2104.03737
| null |
https://arxiv.org/abs/2104.03737v1
|
https://arxiv.org/pdf/2104.03737v1.pdf
|
Few-Shot Action Recognition with Compromised Metric via Optimal Transport
|
Although vital to computer vision systems, few-shot action recognition is still not mature despite the wide research of few-shot image classification. Popular few-shot learning algorithms extract a transferable embedding from seen classes and reuse it on unseen classes by constructing a metric-based classifier. One main obstacle to applying these algorithms in action recognition is the complex structure of videos. Some existing solutions sample frames from a video and aggregate their embeddings to form a video-level representation, neglecting important temporal relations. Others perform an explicit sequence matching between two videos and define their distance as matching cost, imposing too strong restrictions on sequence ordering. In this paper, we propose Compromised Metric via Optimal Transport (CMOT) to combine the advantages of these two solutions. CMOT simultaneously considers semantic and temporal information in videos under Optimal Transport framework, and is discriminative for both content-sensitive and ordering-sensitive tasks. In detail, given two videos, we sample segments from them and cast the calculation of their distance as an optimal transport problem between two segment sequences. To preserve the inherent temporal ordering information, we additionally amend the ground cost matrix by penalizing it with the positional distance between a pair of segments. Empirical results on benchmark datasets demonstrate the superiority of CMOT.
|
['De-Chuan Zhan', 'Han-Jia Ye', 'Su Lu']
|
2021-04-08
| null | null | null | null |
['few-shot-action-recognition']
|
['computer-vision']
|
[ 4.60689098e-01 -4.08388108e-01 -5.68424463e-01 -4.56005126e-01
-5.31804323e-01 -4.38555449e-01 4.88969445e-01 1.06335253e-01
-5.73615432e-01 4.34450418e-01 2.64407277e-01 3.46422493e-01
-4.68382686e-01 -6.09264612e-01 -5.02094388e-01 -8.24753284e-01
-2.08174273e-01 8.34001824e-02 6.48594797e-01 6.55207485e-02
3.87244821e-01 3.77113312e-01 -1.57230604e+00 3.74333173e-01
7.33978868e-01 1.17522478e+00 2.28987753e-01 4.10300046e-01
-1.19364128e-01 1.02991211e+00 -3.06337982e-01 -3.50109518e-01
4.74939138e-01 -8.04829180e-01 -8.54998529e-01 5.38372755e-01
3.15648913e-01 -4.62359041e-01 -6.03984416e-01 1.14558315e+00
2.45115250e-01 5.21108925e-01 5.87158084e-01 -1.53447342e+00
-4.76443470e-01 3.52984458e-01 -4.93808061e-01 5.46053588e-01
4.94303793e-01 1.93576470e-01 1.12906229e+00 -8.15951049e-01
7.84958601e-01 8.79203022e-01 4.17663276e-01 4.10323858e-01
-1.00917935e+00 -8.39870647e-02 1.82800427e-01 1.00187135e+00
-1.29111433e+00 -4.40912247e-01 9.29110944e-01 -6.22925639e-01
7.31596529e-01 3.17879289e-01 8.66453648e-01 8.41618419e-01
1.33522496e-01 9.24786031e-01 6.90201938e-01 -2.62365907e-01
5.07097125e-01 -2.34726697e-01 2.14580074e-01 5.82484305e-01
-3.73685360e-02 -2.63107598e-01 -5.87792695e-01 5.04814126e-02
3.37371707e-01 6.21221423e-01 -5.48291624e-01 -8.95010829e-01
-1.34133077e+00 7.63263106e-01 2.84621537e-01 4.97662574e-01
-2.58429110e-01 5.47516234e-02 6.56896412e-01 4.56963569e-01
3.01774174e-01 1.93930537e-01 -3.99072021e-02 -3.74102056e-01
-8.99594009e-01 2.93592606e-02 4.39506024e-01 9.75898743e-01
9.32513475e-01 -2.54385412e-01 -3.34275573e-01 7.84557104e-01
-1.53789580e-01 3.24739479e-02 7.29275525e-01 -1.02235663e+00
5.58102489e-01 6.44186676e-01 -1.67203620e-01 -1.33719337e+00
3.18091884e-02 1.66742042e-01 -4.75832671e-01 -1.03255160e-01
4.62743133e-01 2.67381728e-01 -6.16894901e-01 1.52541721e+00
3.56888235e-01 5.72254837e-01 -2.54381865e-01 1.11707997e+00
2.87076384e-01 5.62265575e-01 -7.78185651e-02 -5.22856534e-01
1.11365092e+00 -9.83128190e-01 -6.81521893e-01 -7.87491128e-02
8.60565722e-01 -4.56636786e-01 7.97612488e-01 6.41748086e-02
-8.93304646e-01 -4.51507390e-01 -1.25976562e+00 -3.37071940e-02
-3.12289685e-01 -2.09679469e-01 3.27166051e-01 4.21420932e-01
-7.14805841e-01 9.16267097e-01 -8.60528708e-01 -4.85757381e-01
4.47983146e-01 1.22960187e-01 -5.34096360e-01 -3.81585091e-01
-1.08863461e+00 6.27515852e-01 5.38508117e-01 -8.89261365e-02
-7.31083930e-01 -5.28282523e-01 -1.03369725e+00 -1.07934028e-01
7.47100353e-01 -2.82033086e-01 8.85259569e-01 -1.06741452e+00
-1.30523288e+00 6.54320598e-01 -1.20355748e-01 -5.17555177e-01
5.25989532e-01 -8.69845748e-02 -5.29855192e-01 6.88519418e-01
2.19853520e-01 3.36459100e-01 1.10514390e+00 -6.88889921e-01
-7.65407324e-01 -3.60981941e-01 3.37889463e-01 3.47669661e-01
-6.33723855e-01 -1.13480221e-02 -5.09552479e-01 -5.45231700e-01
3.22394282e-01 -8.51447880e-01 -1.23953603e-01 4.63893652e-01
4.55117691e-03 -2.09683001e-01 1.07076681e+00 -4.18140918e-01
1.29529691e+00 -2.37960720e+00 4.84470814e-01 -9.34287682e-02
1.54080421e-01 2.47075826e-01 -2.23602653e-01 5.71080506e-01
4.05142158e-02 -3.41318905e-01 -5.25616050e-01 -4.51729186e-02
-8.83909166e-02 4.44780111e-01 -1.54131919e-01 8.24977517e-01
2.32068002e-01 7.46359348e-01 -1.29132724e+00 -7.98000455e-01
4.83321786e-01 2.78001785e-01 -5.41414618e-01 1.43056169e-01
-4.28313315e-02 1.74537450e-01 -6.20724559e-01 6.13056540e-01
4.79321033e-01 -1.19150719e-02 1.74708247e-01 -3.47738057e-01
-7.70084113e-02 -1.50922209e-01 -1.22098124e+00 2.02951241e+00
-5.95437661e-02 7.50889421e-01 -3.58688146e-01 -1.56202650e+00
7.13166654e-01 1.20464467e-01 1.05864167e+00 -6.30633831e-01
1.10229075e-01 1.07680365e-01 -1.60667449e-01 -9.32258725e-01
3.39077175e-01 -1.66386873e-01 3.12745087e-02 3.27439904e-01
2.24224523e-01 2.71215320e-01 4.62210685e-01 2.75123477e-01
1.30715895e+00 2.43469805e-01 4.53699231e-01 2.11838121e-03
6.44872606e-01 -1.76726684e-01 7.75879562e-01 4.32165354e-01
-7.45436251e-01 8.03668976e-01 4.04141754e-01 -4.98488367e-01
-7.98233271e-01 -9.63187695e-01 4.30369042e-02 1.06673193e+00
5.10153711e-01 -6.49819911e-01 -7.39330173e-01 -9.30165350e-01
-2.43979879e-02 3.49649608e-01 -6.71743870e-01 -3.91288310e-01
-7.74784207e-01 -4.06708658e-01 1.76356226e-01 4.48203862e-01
5.14120698e-01 -7.64538467e-01 -9.55895603e-01 3.02621037e-01
-3.51374477e-01 -1.14891446e+00 -8.12158406e-01 -8.02476555e-02
-7.22222269e-01 -1.29243875e+00 -8.80896688e-01 -7.72033572e-01
5.25786221e-01 7.86216736e-01 5.72864234e-01 -8.66082087e-02
-5.48628449e-01 5.60571849e-01 -7.32668340e-01 1.67109847e-01
9.39774737e-02 -3.89062464e-01 1.10772103e-01 7.22441852e-01
5.29122710e-01 -6.22359753e-01 -6.75453544e-01 5.25667131e-01
-9.98463213e-01 -3.01824778e-01 3.15261513e-01 7.06626892e-01
6.33356750e-01 2.72934027e-02 2.69049376e-01 -4.02314514e-01
6.33726493e-02 -4.95810717e-01 -1.02261119e-01 3.42977941e-01
-2.62586147e-01 -4.56927009e-02 5.39712369e-01 -4.83430654e-01
-6.86809897e-01 1.94354087e-01 4.07745212e-01 -8.88229370e-01
1.04101166e-01 3.32671613e-01 -3.70044798e-01 -3.13710980e-02
1.87695593e-01 4.66489732e-01 1.17418244e-01 -2.49020517e-01
4.10451531e-01 5.34483850e-01 3.79648715e-01 -3.19652796e-01
6.28192782e-01 7.92176008e-01 -7.46150538e-02 -1.04467368e+00
-8.51558685e-01 -8.93444180e-01 -1.06548810e+00 -5.21324515e-01
1.05762899e+00 -4.41276789e-01 -3.68996739e-01 3.54971707e-01
-8.61654341e-01 1.33608848e-01 -5.84121287e-01 7.92592108e-01
-8.08752000e-01 8.28161061e-01 -2.94974953e-01 -5.81073761e-01
2.11580366e-01 -1.04723203e+00 8.70061219e-01 -4.96672690e-02
-9.68024433e-02 -8.53936613e-01 2.70218402e-01 2.32828945e-01
-7.63716847e-02 3.27314615e-01 6.63641036e-01 -5.00051618e-01
-5.98273695e-01 -3.21001798e-01 -1.17464244e-01 3.73610675e-01
3.17113191e-01 -7.73465782e-02 -6.91428900e-01 -3.68153065e-01
3.92950594e-01 -1.13893569e-01 1.02533805e+00 2.70467997e-01
1.18535662e+00 -2.78142929e-01 -3.50623131e-01 5.97303569e-01
1.48683798e+00 3.66669565e-01 6.22831762e-01 2.66931891e-01
8.31851482e-01 7.48446822e-01 8.11229110e-01 5.02535522e-01
9.59773213e-02 8.01828802e-01 4.33854520e-01 6.79423690e-01
-4.29705642e-02 -2.19300389e-01 5.80924571e-01 1.03901350e+00
-1.46620139e-01 -2.85950135e-02 -5.25785625e-01 5.40891111e-01
-2.08707213e+00 -1.45205569e+00 1.70725122e-01 2.25783801e+00
5.61773241e-01 1.24190152e-01 1.56821162e-01 4.07675296e-01
8.76338601e-01 6.23218238e-01 -5.06995380e-01 -1.05122574e-01
-4.09103371e-02 -2.30749488e-01 3.82309198e-01 9.32453945e-02
-1.21130359e+00 5.00911534e-01 5.30436754e+00 9.87200618e-01
-9.09054995e-01 1.98417634e-01 1.78957090e-01 -3.98554802e-01
9.91663113e-02 1.79052204e-01 -4.30435449e-01 7.38281786e-01
7.02342391e-01 -3.92852604e-01 2.57225573e-01 8.18015337e-01
1.76337063e-01 -1.49906725e-01 -1.46808445e+00 1.13718557e+00
5.03975213e-01 -1.21519864e+00 1.81408271e-01 2.07361132e-02
5.18530369e-01 -2.51398236e-01 -1.10998169e-01 1.49445459e-01
-4.22653288e-01 -5.58266103e-01 7.75662959e-01 5.02427936e-01
5.25755346e-01 -6.17233157e-01 4.53438133e-01 1.20950714e-01
-1.46047831e+00 -2.53216803e-01 -6.20403707e-01 -1.56880066e-01
2.21896201e-01 3.19763809e-01 -2.33132392e-01 7.47231781e-01
5.49559176e-01 1.55109680e+00 -5.38805485e-01 1.24130452e+00
2.52003044e-01 2.49910876e-01 2.35026646e-02 6.05376624e-02
5.08431315e-01 -6.28339946e-01 7.83933580e-01 9.27296996e-01
2.75724113e-01 2.03775510e-01 4.04126316e-01 3.58737528e-01
1.67768776e-01 2.00656310e-01 -8.50101471e-01 -6.06612526e-02
2.37018839e-01 1.04825592e+00 -8.85649145e-01 -4.11970377e-01
-7.00790286e-01 1.36227643e+00 2.69773304e-01 1.48677707e-01
-9.38291132e-01 -4.89371687e-01 7.78928459e-01 -4.43228008e-03
5.06654859e-01 -1.76622272e-01 1.56729981e-01 -1.34398174e+00
3.82347792e-01 -5.99156976e-01 6.39189184e-01 -4.39114124e-01
-1.03968608e+00 2.26259977e-01 1.56927422e-01 -2.03223252e+00
5.14217652e-02 -4.54611987e-01 -4.93567497e-01 -4.04570764e-03
-1.31049418e+00 -7.64108539e-01 -3.10910940e-01 7.76474237e-01
1.11206806e+00 -9.76078808e-02 4.23268408e-01 5.31727195e-01
-6.48230910e-01 3.64069253e-01 2.69559771e-01 2.01165304e-01
6.20711207e-01 -9.68656361e-01 -1.22652657e-01 8.72363687e-01
3.25573891e-01 3.50623459e-01 5.35929382e-01 -4.54934835e-01
-1.58250916e+00 -1.19824171e+00 7.05295026e-01 -3.72291863e-01
8.72858942e-01 -1.72998533e-01 -1.03194976e+00 5.50156355e-01
-6.88378839e-03 3.07615846e-01 6.43129528e-01 -4.36036974e-01
-4.54709202e-01 -2.41691098e-01 -9.30219710e-01 5.58257699e-01
1.43004346e+00 -6.23619556e-01 -7.88871169e-01 3.67558718e-01
6.11191392e-01 1.96773902e-01 -7.39434779e-01 2.18822271e-01
5.96666992e-01 -1.17699671e+00 8.21135283e-01 -7.47874498e-01
3.75390470e-01 -4.58226055e-01 -4.94352579e-01 -1.18307710e+00
-2.96859354e-01 -5.88788331e-01 -1.79450348e-01 9.85840082e-01
-2.00093403e-01 -4.16342646e-01 6.52717352e-01 4.02478993e-01
-1.61811575e-01 -6.93847001e-01 -1.12313724e+00 -1.27767158e+00
-3.34667206e-01 -3.64518553e-01 3.94823402e-01 1.07923901e+00
3.38004142e-01 3.72222252e-03 -5.86324513e-01 -5.12555093e-02
8.07421505e-01 2.01501742e-01 4.40329492e-01 -9.83226895e-01
-1.95584387e-01 -4.46350217e-01 -1.10019410e+00 -9.35672462e-01
1.60438225e-01 -8.61721039e-01 4.39194851e-02 -1.24505818e+00
3.34232897e-01 4.69666831e-02 -5.64111352e-01 1.40863776e-01
8.30084905e-02 3.01967949e-01 3.33388716e-01 4.46209550e-01
-9.11429524e-01 8.31715107e-01 1.08444846e+00 -4.33481127e-01
1.92389134e-02 -2.92047679e-01 -4.42718202e-03 8.00006628e-01
5.57012379e-01 -4.31765109e-01 -7.49498665e-01 -3.45034629e-01
-3.65928441e-01 7.32347295e-02 3.58205676e-01 -1.29499865e+00
2.71085471e-01 -4.16209251e-01 7.27731735e-02 -3.73627096e-01
4.78711218e-01 -9.71045852e-01 1.10140312e-02 5.00680029e-01
-4.31396723e-01 -2.69506961e-01 -4.44531739e-01 9.92854238e-01
-4.28752363e-01 -4.63994801e-01 8.72488081e-01 -2.45452039e-02
-1.31864631e+00 5.96866548e-01 -1.41618207e-01 1.49424270e-01
1.63341641e+00 -7.90884614e-01 -2.30228137e-02 -1.00500517e-01
-6.83702588e-01 7.65153840e-02 5.46202362e-01 6.03097260e-01
8.66914213e-01 -1.59979749e+00 -3.02356690e-01 2.66830683e-01
4.92881030e-01 -4.98072714e-01 3.99059862e-01 1.08369696e+00
-3.59234273e-01 2.63777018e-01 -4.95641857e-01 -6.30860388e-01
-1.28127754e+00 8.54719579e-01 2.16997951e-01 2.44733527e-01
-1.02009606e+00 5.52883267e-01 2.16840461e-01 2.89476007e-01
2.07141295e-01 -3.33236992e-01 -3.15251321e-01 4.93604302e-01
7.72570908e-01 6.45461798e-01 -1.95806757e-01 -8.67634952e-01
-4.90488112e-01 7.83460617e-01 -8.03917795e-02 6.76011741e-02
1.19863594e+00 -2.83732384e-01 1.27375826e-01 7.46870637e-01
1.68288493e+00 -5.42014778e-01 -1.47950089e+00 -3.41668814e-01
2.95028001e-01 -1.02585447e+00 -1.71665534e-01 5.21751978e-02
-1.11573887e+00 8.46687794e-01 3.39187413e-01 3.11825216e-01
1.02327895e+00 -1.05049103e-01 1.00033557e+00 3.06807846e-01
4.59643275e-01 -1.31289911e+00 5.48650801e-01 2.38844812e-01
6.18953645e-01 -1.15197873e+00 3.71617712e-02 -3.41094673e-01
-6.57094061e-01 1.05165100e+00 3.95533711e-01 -2.39914149e-01
6.01694643e-01 -2.79122084e-01 -3.94230932e-01 -1.28510654e-01
-5.07976532e-01 -2.88064450e-01 2.34947428e-01 5.76306999e-01
-5.50902821e-02 -1.49399444e-01 -5.18178165e-01 1.22045413e-01
4.40018207e-01 9.62524861e-02 5.53213656e-01 1.30001664e+00
-7.04152524e-01 -8.78715515e-01 1.68428253e-02 4.56774920e-01
-1.98109195e-01 2.05779970e-01 -1.45909697e-01 4.85424340e-01
8.20114538e-02 8.01010013e-01 1.74753785e-01 -5.43083072e-01
2.87728369e-01 9.78440419e-02 5.38103878e-01 -6.40995502e-01
6.47706017e-02 -5.95043339e-02 -1.70370802e-01 -8.97842050e-01
-7.36385465e-01 -1.01487899e+00 -1.05567586e+00 1.07737994e-02
-1.90958813e-01 1.23818286e-01 2.91685045e-01 1.04932201e+00
2.43829727e-01 1.90645486e-01 9.80442882e-01 -7.87800014e-01
-7.08969176e-01 -5.39232492e-01 -9.11180556e-01 8.15736353e-01
3.70535880e-01 -9.99356151e-01 -5.41081965e-01 2.99369544e-01]
|
[8.48694133758545, 0.7433567643165588]
|
c6d94925-70a7-4903-a694-7175ba3e4cd4
|
joint-aspect-extraction-and-sentiment
| null | null |
https://aclanthology.org/2020.coling-main.24
|
https://aclanthology.org/2020.coling-main.24.pdf
|
Joint Aspect Extraction and Sentiment Analysis with Directional Graph Convolutional Networks
|
End-to-end aspect-based sentiment analysis (EASA) consists of two sub-tasks: the first extracts the aspect terms in a sentence and the second predicts the sentiment polarities for such terms. For EASA, compared to pipeline and multi-task approaches, joint aspect extraction and sentiment analysis provides a one-step solution to predict both aspect terms and their sentiment polarities through a single decoding process, which avoid the mismatches in between the results of aspect terms and sentiment polarities, as well as error propagation. Previous studies, especially recent ones, for this task focus on using powerful encoders (e.g., Bi-LSTM and BERT) to model contextual information from the input, with limited efforts paid to using advanced neural architectures (such as attentions and graph convolutional networks) or leveraging extra knowledge (such as syntactic information). To extend such efforts, in this paper, we propose directional graph convolutional networks (D-GCN) to jointly perform aspect extraction and sentiment analysis with encoding syntactic information, where dependency among words are integrated in our model to enhance its ability of representing input sentences and help EASA accordingly. Experimental results on three benchmark datasets demonstrate the effectiveness of our approach, where D-GCN achieves state-of-the-art performance on all datasets.
|
['Yan Song', 'Yuanhe Tian', 'Guimin Chen']
|
2020-12-01
| null | null | null |
coling-2020-8
|
['aspect-extraction']
|
['natural-language-processing']
|
[ 2.21781656e-01 3.74132633e-01 -1.49421141e-01 -7.58787453e-01
-6.08987451e-01 -6.16339743e-01 6.38286471e-01 2.26535544e-01
-2.98326463e-01 1.88210398e-01 5.48509240e-01 -5.12453139e-01
3.00594866e-01 -9.47951615e-01 -6.14975870e-01 -4.17270899e-01
5.66349700e-02 2.82133728e-01 -1.67502135e-01 -5.82443058e-01
2.21046686e-01 -1.15030527e-01 -1.02669954e+00 5.77224791e-01
7.95450866e-01 1.25875401e+00 -1.02037527e-01 6.04564011e-01
-6.19052649e-01 9.14775491e-01 -5.35877883e-01 -9.81854022e-01
-7.39550889e-02 -2.43807450e-01 -7.94631779e-01 1.40363667e-02
-1.13232888e-01 1.20550040e-02 1.37480870e-02 1.04069281e+00
3.15549910e-01 3.06733977e-02 5.63477635e-01 -9.51477826e-01
-9.73573148e-01 1.03098404e+00 -6.10950828e-01 -1.88331455e-01
1.74809068e-01 1.65948659e-01 1.67991352e+00 -9.53509748e-01
2.96783000e-01 1.20314598e+00 8.37496817e-01 2.86986560e-01
-6.60359681e-01 -3.53663802e-01 8.27017426e-01 5.82294613e-02
-7.89118350e-01 -3.25044006e-01 1.24882805e+00 -1.99652001e-01
1.58313239e+00 1.86131373e-01 8.52516651e-01 1.07397139e+00
4.41268384e-01 1.22920442e+00 8.81368637e-01 -1.31883278e-01
-2.05563176e-02 3.45178172e-02 4.58482921e-01 7.07732081e-01
1.82356536e-01 -5.78219712e-01 -5.67258060e-01 1.01122990e-01
2.92099882e-02 -1.33628026e-01 5.67340739e-02 1.43468902e-01
-9.76580858e-01 9.33349490e-01 6.79084599e-01 1.81373864e-01
-7.15878069e-01 7.57177621e-02 6.52011096e-01 2.56176174e-01
1.04115403e+00 4.86968875e-01 -8.17970932e-01 -2.84062885e-03
-7.03128695e-01 -2.68355906e-02 7.84625173e-01 1.03608406e+00
7.62016952e-01 1.34382412e-01 -4.91527200e-01 7.10363090e-01
6.84817970e-01 5.42732835e-01 4.31488991e-01 7.52825215e-02
8.76082063e-01 1.22840726e+00 -4.04435426e-01 -1.09571314e+00
-6.98200881e-01 -5.75377524e-01 -8.15218627e-01 -2.11530134e-01
-1.41885579e-01 -4.19585377e-01 -1.24829614e+00 1.61211586e+00
2.04687312e-01 -2.85558134e-01 2.04937145e-01 6.59492016e-01
1.23500597e+00 7.25562274e-01 1.90386802e-01 6.82727993e-02
1.71580315e+00 -1.34953177e+00 -8.47874045e-01 -9.69289362e-01
8.34183395e-01 -8.65323424e-01 1.10508621e+00 1.38094038e-01
-1.03803122e+00 -2.91367292e-01 -1.02790964e+00 -3.52133125e-01
-7.44994640e-01 1.50401220e-01 1.03157580e+00 3.46672595e-01
-1.10259128e+00 3.62405062e-01 -8.59700799e-01 -2.52398849e-02
7.39866734e-01 4.77111101e-01 -1.70180723e-01 1.16515517e-01
-1.37787068e+00 7.26919353e-01 2.20792532e-01 6.03123426e-01
-3.70394856e-01 -5.36279261e-01 -1.36227489e+00 2.76602983e-01
2.95638978e-01 -1.02394545e+00 1.08843267e+00 -1.23402202e+00
-1.53804696e+00 6.27412021e-01 -4.17689979e-01 -1.79763094e-01
-1.39463425e-01 -3.38734627e-01 -3.31015348e-01 -2.45824486e-01
8.93306807e-02 3.80363524e-01 6.66875482e-01 -1.12886584e+00
-4.38040107e-01 -6.06328011e-01 6.20515585e-01 4.13926482e-01
-6.86267257e-01 3.13906729e-01 -7.60790229e-01 -6.02622628e-01
-9.13372189e-02 -9.35407639e-01 -4.84867990e-01 -5.91754436e-01
-8.94790471e-01 -4.79164183e-01 6.40425861e-01 -7.74031520e-01
1.36416030e+00 -1.80834770e+00 2.45577395e-01 1.09445982e-01
1.80965781e-01 3.46288234e-01 -4.13983285e-01 4.47140545e-01
4.26350422e-02 2.31386185e-01 -3.39967847e-01 -9.02207315e-01
2.23976105e-01 7.73793906e-02 -3.13348502e-01 2.84773353e-02
7.44022012e-01 1.37632549e+00 -7.80641258e-01 -3.14501047e-01
-1.96848467e-01 7.27635562e-01 -6.18206620e-01 2.68341124e-01
-5.06790698e-01 1.16056554e-01 -6.44752741e-01 8.42861235e-01
6.06716931e-01 -4.86509025e-01 2.07093075e-01 -3.45180959e-01
1.50809973e-01 9.89494205e-01 -6.78185105e-01 1.46074235e+00
-9.76810217e-01 4.30550098e-01 -1.97051223e-02 -8.78705621e-01
8.97071838e-01 2.45915607e-01 1.48222268e-01 -7.67040253e-01
4.65418994e-01 6.52304664e-03 -2.87910532e-02 -3.83845657e-01
7.61343658e-01 -1.66092679e-01 -4.04358774e-01 4.55656379e-01
2.94675827e-01 -1.19891375e-01 2.82324135e-01 2.71293044e-01
8.24839771e-01 8.80004391e-02 2.74607390e-01 -1.56714201e-01
7.68468440e-01 -3.48729119e-02 5.48076451e-01 3.22867870e-01
8.97749066e-02 4.64863986e-01 1.06477845e+00 -4.23960030e-01
-7.36442089e-01 -3.80813599e-01 3.54896933e-01 1.21964622e+00
2.68438533e-02 -8.39173436e-01 -5.55827856e-01 -1.13848543e+00
-2.66000688e-01 7.13372409e-01 -8.96291494e-01 -9.43911970e-02
-7.08149850e-01 -1.10787237e+00 1.99359134e-01 7.25882888e-01
4.13406253e-01 -1.18835473e+00 9.62868333e-02 1.03829138e-01
-1.63715303e-01 -1.32300246e+00 -3.38558167e-01 4.42324579e-01
-5.98684549e-01 -8.24591219e-01 -3.60596120e-01 -8.09844494e-01
7.31886744e-01 1.48522899e-01 1.34418488e+00 1.28840497e-02
4.44481492e-01 -6.89506456e-02 -5.66940904e-01 -6.38815701e-01
-7.51339346e-02 4.73551244e-01 -4.67323601e-01 1.87271252e-01
6.37955010e-01 -5.72208166e-01 -8.09402108e-01 -1.48161188e-01
-8.35524201e-01 2.49080822e-01 8.48782063e-01 8.10629666e-01
6.55239224e-01 -2.61835665e-01 5.29016554e-01 -1.55619013e+00
9.68286812e-01 -6.02266490e-01 -2.29222149e-01 1.42149314e-01
-7.83425331e-01 -1.19390294e-01 1.10649145e+00 2.09635962e-02
-1.08304572e+00 -3.11292470e-01 -5.05543411e-01 5.89870755e-03
2.05563337e-01 1.23261845e+00 -3.78990561e-01 2.32513323e-01
4.46077920e-02 2.74487644e-01 -3.57077837e-01 -1.58907712e-01
5.82824707e-01 7.29611933e-01 4.53001671e-02 -7.29733780e-02
4.31523144e-01 3.59163284e-01 -9.20824558e-02 -4.57870543e-01
-1.55295432e+00 -4.56046343e-01 -4.72544402e-01 -1.28726453e-01
8.24331641e-01 -1.22514570e+00 -4.45655167e-01 7.07168400e-01
-1.28802884e+00 -3.67790759e-02 1.53219169e-02 1.68925986e-01
-1.38737604e-01 2.43951112e-01 -6.35404766e-01 -6.47597551e-01
-1.10586309e+00 -1.32795537e+00 1.34851801e+00 2.75506437e-01
-2.50764728e-01 -1.40067971e+00 -1.19678505e-01 6.91177845e-01
6.91302717e-01 3.24015692e-02 9.99198556e-01 -9.18326616e-01
-3.80905032e-01 -3.48871499e-01 -2.05992460e-01 3.97433460e-01
1.70500860e-01 -1.38234749e-01 -1.20116282e+00 -1.59330487e-01
-5.16683534e-02 -3.41384828e-01 1.03868461e+00 3.15651745e-01
9.45957541e-01 -5.75801730e-01 -1.03440084e-01 7.43066490e-01
1.24319875e+00 -8.38056505e-02 4.60029811e-01 5.53970277e-01
1.35506117e+00 5.56787848e-01 5.44050157e-01 2.62575388e-01
1.13325274e+00 3.10110182e-01 5.99505544e-01 -2.76058882e-01
-2.05552861e-01 -1.78873882e-01 6.44630671e-01 1.59806490e+00
6.58481568e-02 -5.87553442e-01 -7.65284896e-01 8.22075665e-01
-1.82003915e+00 -3.29286635e-01 -3.92594606e-01 1.43898547e+00
8.04435134e-01 3.62203121e-01 -1.63166866e-01 -8.44135061e-02
2.12719649e-01 9.77325797e-01 -5.93685031e-01 -9.00354326e-01
-3.23156625e-01 2.14693129e-01 2.90703714e-01 5.44813514e-01
-1.38065243e+00 1.28728426e+00 5.09752989e+00 6.76308990e-01
-1.25365126e+00 2.32012440e-02 8.72737527e-01 -2.99242157e-02
-9.06697512e-01 6.62252754e-02 -6.69110060e-01 2.91765481e-01
9.27184165e-01 1.11267686e-01 1.73161536e-01 9.58837450e-01
-3.55234407e-02 2.65333444e-01 -7.68438041e-01 4.13620144e-01
3.24355274e-01 -1.05586135e+00 3.81083965e-01 -3.99686992e-01
9.02406275e-01 2.62077868e-01 1.08640067e-01 6.18951917e-01
3.76795828e-01 -8.69448125e-01 5.99675655e-01 2.22410411e-01
4.38424617e-01 -8.70544016e-01 1.13884604e+00 -7.49576315e-02
-1.35384893e+00 1.02927521e-01 -2.51869112e-01 -2.44118229e-01
3.30452800e-01 9.37748313e-01 -6.58493042e-01 9.04426336e-01
5.57003498e-01 1.15020633e+00 -6.52719378e-01 3.80686045e-01
-8.41901541e-01 7.48594344e-01 -5.27499095e-02 -5.15749574e-01
6.50677323e-01 -2.92689800e-01 4.94163364e-01 1.46855712e+00
2.73338389e-02 -2.06857175e-01 -1.43445984e-01 6.84784770e-01
-3.45171213e-01 4.24298674e-01 -4.92442966e-01 -4.28446174e-01
-1.23387642e-01 1.72004640e+00 -7.42924154e-01 -3.66534680e-01
-8.51144314e-01 7.40199327e-01 6.59668028e-01 4.57618117e-01
-5.24044573e-01 -6.32181346e-01 8.61751676e-01 -2.71983027e-01
6.80122077e-01 -1.29823029e-01 -9.35330749e-01 -1.34935045e+00
3.52979600e-01 -8.19800913e-01 4.35758621e-01 -6.27276719e-01
-1.20322025e+00 1.13749087e+00 -5.56848347e-01 -1.04512763e+00
-2.65766948e-01 -7.43485808e-01 -9.86425340e-01 1.04455817e+00
-2.15740585e+00 -1.86993849e+00 2.08033156e-02 3.83056402e-01
4.96626019e-01 -5.11881113e-02 6.70974195e-01 1.85135424e-01
-6.26035869e-01 6.39487445e-01 -4.27046388e-01 3.06979805e-01
3.25007200e-01 -1.38377583e+00 1.04968691e+00 9.75363612e-01
3.78790195e-03 6.92221582e-01 4.93165791e-01 -6.38270080e-01
-1.80292058e+00 -1.25147200e+00 1.46882141e+00 -4.39971805e-01
1.00272787e+00 -6.79066598e-01 -7.81338573e-01 8.60780716e-01
5.29726803e-01 -1.54185116e-01 8.69721532e-01 7.62273490e-01
-3.79684269e-01 4.74103056e-02 -4.28255916e-01 6.83803916e-01
8.31919253e-01 -7.24546731e-01 -5.08225560e-01 2.82914937e-01
1.02297258e+00 -3.27491850e-01 -7.22218454e-01 6.51535749e-01
2.16564149e-01 -8.40195239e-01 6.43172443e-01 -7.54496515e-01
9.90726113e-01 -1.72605246e-01 -2.66292226e-02 -1.71536291e+00
-2.68330783e-01 -2.89602131e-01 -2.09770232e-01 1.47963977e+00
9.81739104e-01 -5.66697657e-01 5.27830422e-01 4.82822031e-01
-4.29307491e-01 -1.49793410e+00 -4.40175503e-01 -8.56359676e-02
-4.33037570e-03 -6.83920145e-01 7.70717144e-01 8.46299529e-01
6.57410398e-02 9.56602752e-01 -2.76278079e-01 3.20419490e-01
1.01228394e-01 6.51933670e-01 5.01422942e-01 -8.13554406e-01
-1.60304919e-01 -7.14575827e-01 -1.51215211e-01 -9.94343698e-01
4.37704355e-01 -1.12678158e+00 9.60568525e-03 -2.17605639e+00
1.95000470e-01 -1.55978888e-01 -4.43662614e-01 7.10720181e-01
-8.58021855e-01 8.71005505e-02 7.60357901e-02 -2.56183833e-01
-9.95301247e-01 9.84985590e-01 1.38889909e+00 -4.50050145e-01
-9.66581479e-02 7.00815842e-02 -1.44726634e+00 9.16786015e-01
6.18729889e-01 -4.29215312e-01 -4.20221835e-01 -8.88905823e-01
9.81137335e-01 -2.65122980e-01 -2.18450889e-01 -3.27890962e-01
2.56824583e-01 1.78812981e-01 1.50984079e-01 -6.37919843e-01
3.15905392e-01 -5.40415585e-01 -5.98699093e-01 5.19068446e-03
-2.46703267e-01 1.79164916e-01 1.52687281e-01 4.30295020e-01
-7.31420994e-01 -2.22406257e-02 1.09567456e-01 -4.89317030e-02
-5.92934012e-01 4.46317494e-01 -1.09608687e-01 1.00823276e-01
5.23385942e-01 3.21412772e-01 -5.07999122e-01 -4.46379215e-01
-3.31375539e-01 5.32980919e-01 7.47312531e-02 5.91253638e-01
5.03188848e-01 -1.19387913e+00 -6.73173070e-01 2.10550949e-01
2.18633994e-01 4.06318754e-01 3.79521459e-01 8.36007297e-01
-1.39715523e-01 5.13123810e-01 3.46047133e-01 -2.32559383e-01
-1.10246158e+00 4.29978758e-01 7.62494355e-02 -8.53264093e-01
-2.90186197e-01 1.20725727e+00 2.63083726e-01 -8.06285024e-01
-1.79785952e-01 -4.51727748e-01 -7.74211526e-01 4.43936706e-01
4.01874155e-01 -4.15591538e-01 3.32735717e-01 -7.35640347e-01
-5.38490534e-01 6.38288140e-01 -3.46060991e-01 2.85288453e-01
1.62869275e+00 -1.70690238e-01 -4.26600963e-01 3.47406417e-01
1.28281593e+00 8.19839463e-02 -9.25493300e-01 -3.34785134e-01
-2.45422516e-02 -5.17192893e-02 3.11328381e-01 -9.13962483e-01
-1.40912521e+00 1.02767646e+00 -3.28526646e-01 4.25066650e-01
1.34020340e+00 9.26493779e-02 1.10620701e+00 3.29295218e-01
-1.88042477e-01 -9.52646911e-01 -9.94992852e-02 9.69741523e-01
7.93310761e-01 -1.46815300e+00 5.70875965e-02 -3.70342404e-01
-1.08936262e+00 9.70034540e-01 6.17797256e-01 -2.01582521e-01
7.21401095e-01 2.63396174e-01 4.15609419e-01 -5.33186257e-01
-1.06695652e+00 -3.96528631e-01 5.97157419e-01 1.73532963e-01
7.07897604e-01 1.82495460e-01 -3.06587011e-01 1.34968996e+00
-4.60374713e-01 -3.70134115e-01 2.27768183e-01 8.33355784e-01
1.21857980e-02 -9.55877006e-01 3.86543959e-01 6.31008029e-01
-8.48142326e-01 -7.38722682e-01 -5.13452828e-01 3.66467655e-01
-2.50291765e-01 1.05923998e+00 -7.62965605e-02 -6.19489312e-01
5.95499456e-01 1.12636752e-01 -1.22141354e-01 -5.77830970e-01
-1.15103459e+00 1.01068757e-01 6.45454466e-01 -6.03159487e-01
-5.35120308e-01 -4.64420885e-01 -1.15492737e+00 6.55654701e-04
-3.01595956e-01 -9.41534862e-02 1.02389288e+00 1.24775410e+00
5.92515886e-01 1.13867116e+00 8.77299666e-01 -5.74143946e-01
-9.59196240e-02 -1.19692099e+00 -3.39480042e-01 3.83737594e-01
2.99678564e-01 -1.02492318e-01 -1.74610585e-01 -3.17754865e-01]
|
[11.47685718536377, 6.670983791351318]
|
f6a1293f-cc55-49ea-995f-6756c6331076
|
robust-submodular-maximization-a-non-uniform
|
1706.04918
| null |
http://arxiv.org/abs/1706.04918v1
|
http://arxiv.org/pdf/1706.04918v1.pdf
|
Robust Submodular Maximization: A Non-Uniform Partitioning Approach
|
We study the problem of maximizing a monotone submodular function subject to
a cardinality constraint $k$, with the added twist that a number of items
$\tau$ from the returned set may be removed. We focus on the worst-case setting
considered in (Orlin et al., 2016), in which a constant-factor approximation
guarantee was given for $\tau = o(\sqrt{k})$. In this paper, we solve a key
open problem raised therein, presenting a new Partitioned Robust (PRo)
submodular maximization algorithm that achieves the same guarantee for more
general $\tau = o(k)$. Our algorithm constructs partitions consisting of
buckets with exponentially increasing sizes, and applies standard submodular
optimization subroutines on the buckets in order to construct the robust
solution. We numerically demonstrate the performance of PRo in data
summarization and influence maximization, demonstrating gains over both the
greedy algorithm and the algorithm of (Orlin et al., 2016).
|
['Volkan Cevher', 'Slobodan Mitrović', 'Jonathan Scarlett', 'Ilija Bogunovic']
|
2017-06-15
|
robust-submodular-maximization-a-non-uniform-1
|
https://icml.cc/Conferences/2017/Schedule?showEvent=674
|
http://proceedings.mlr.press/v70/bogunovic17a/bogunovic17a.pdf
|
icml-2017-8
|
['data-summarization']
|
['miscellaneous']
|
[ 2.55367488e-01 4.90745425e-01 -4.50411677e-01 -1.51555955e-01
-9.58090007e-01 -1.00276828e+00 -3.30735356e-01 5.17303526e-01
-3.89294028e-01 9.46023941e-01 4.30462286e-02 -5.69506697e-02
-7.02235460e-01 -8.67053390e-01 -9.70418334e-01 -8.60230207e-01
-5.66135764e-01 7.14975893e-01 -1.08565725e-01 -1.35141656e-01
3.89543861e-01 2.94893652e-01 -1.29855990e+00 -7.71625713e-02
9.10053432e-01 1.13841319e+00 1.17061540e-01 4.78820354e-01
1.16107307e-01 3.40627044e-01 -5.29811323e-01 -2.56623685e-01
8.75654876e-01 -2.06551895e-01 -8.49884927e-01 5.64514697e-01
5.47788739e-01 -2.63465494e-01 -3.47341627e-01 1.14977729e+00
3.51317048e-01 5.20023629e-02 3.76046926e-01 -1.48186541e+00
-1.74778834e-01 1.09800589e+00 -1.14641523e+00 1.93032660e-02
2.46761546e-01 -2.02673197e-01 1.39443600e+00 -6.07725501e-01
7.68215597e-01 1.00897956e+00 1.58041164e-01 1.12233378e-01
-1.36372995e+00 -7.67100573e-01 3.71199846e-01 -9.85917822e-02
-1.57597601e+00 -3.38141382e-01 5.94860673e-01 -7.49726444e-02
7.51260877e-01 7.27221847e-01 6.13493264e-01 -2.32850850e-01
-1.00568242e-01 9.63787973e-01 7.08124161e-01 -2.05335528e-01
3.52277547e-01 9.34045687e-02 3.15411448e-01 3.80507022e-01
1.03377116e+00 -5.03290176e-01 -6.74553990e-01 -6.64293289e-01
1.17409952e-01 -1.11943834e-01 -4.92390662e-01 -6.88524961e-01
-9.10628617e-01 9.91954207e-01 2.92630315e-01 -1.62609100e-01
-3.24834675e-01 5.67567647e-01 1.54750839e-01 5.25427878e-01
5.23850501e-01 4.48999405e-01 -6.06608689e-01 1.61471143e-01
-1.37299657e+00 8.36187661e-01 1.03882372e+00 1.54856944e+00
7.75989234e-01 -2.31254071e-01 -7.23557100e-02 5.15058875e-01
6.47119358e-02 6.49942815e-01 -4.82091129e-01 -1.31215954e+00
9.64331210e-01 6.91861033e-01 3.62324774e-01 -9.83419895e-01
-4.39327568e-01 -3.26909095e-01 -5.60173690e-01 -1.63016438e-01
3.90376091e-01 -2.94341832e-01 -4.95241493e-01 1.87182975e+00
5.86722612e-01 -5.27547121e-01 -5.01770116e-02 7.07187116e-01
2.62118638e-01 8.11450481e-01 -5.72795451e-01 -1.03972244e+00
9.19932663e-01 -6.38904452e-01 -5.47898293e-01 -1.71669049e-03
9.38636065e-01 -6.39223278e-01 4.81106132e-01 7.73361266e-01
-1.75676954e+00 3.69022727e-01 -1.05915689e+00 -6.16508722e-02
1.36161372e-01 -4.49618310e-01 6.23109162e-01 7.82039046e-01
-1.16941810e+00 2.73947269e-01 -3.82909745e-01 -7.09240288e-02
5.12193501e-01 6.86252058e-01 -6.95278794e-02 -4.48531300e-01
-5.33903062e-01 3.27691138e-01 4.09645706e-01 -1.56088665e-01
-8.41853678e-01 -9.93134737e-01 -7.63275564e-01 1.75548568e-01
1.02387094e+00 -7.58724988e-01 1.05145788e+00 -4.39528048e-01
-7.82843888e-01 8.56882393e-01 -2.28905603e-01 -5.01715899e-01
5.44979036e-01 -1.76690128e-02 5.32801211e-01 3.03434014e-01
8.90964642e-03 4.01137352e-01 5.20528316e-01 -1.30707228e+00
-7.05854177e-01 -8.43106449e-01 5.22138119e-01 2.85601795e-01
-4.97307897e-01 -6.00423589e-02 -3.67336839e-01 -4.31310058e-01
4.09739792e-01 -9.04988229e-01 -6.17246330e-01 -1.29498228e-01
-5.62702358e-01 -2.94728786e-01 3.82710725e-01 -4.17316318e-01
1.58821356e+00 -1.89481020e+00 5.40972948e-01 6.78057313e-01
5.72509050e-01 -3.85412425e-01 3.87878194e-02 7.39037395e-01
2.94719368e-01 2.36957118e-01 -7.74185419e-01 -3.36873055e-01
1.68471619e-01 1.64138168e-01 -8.64768997e-02 9.49063480e-01
-6.03977203e-01 6.49505496e-01 -5.18641710e-01 -2.09196702e-01
-5.06619453e-01 -4.74598110e-01 -8.10423613e-01 -2.35176966e-01
-7.38902628e-01 -3.97739530e-01 -2.28063911e-01 9.78165925e-01
1.27321136e+00 -1.59981921e-01 5.41867197e-01 1.81376055e-01
-1.68150038e-01 -1.31921157e-01 -1.77091610e+00 1.61643457e+00
-1.41733199e-01 1.88398153e-01 8.76131535e-01 -1.07522655e+00
6.66548133e-01 -2.37849485e-02 9.56949532e-01 -3.80712152e-01
1.77849486e-01 4.14275676e-01 -2.71885961e-01 -1.78784072e-01
8.65871966e-01 -1.16110161e-01 -4.83609021e-01 9.61178064e-01
-4.96103138e-01 -2.11346403e-01 6.26094103e-01 7.69479156e-01
1.15893269e+00 -8.13845098e-01 2.61956900e-01 -7.28235781e-01
2.08611935e-01 2.50565857e-01 6.19240463e-01 7.61583209e-01
1.85144708e-01 5.31994283e-01 1.06022072e+00 -1.00961337e-02
-1.13834095e+00 -9.17721808e-01 -2.90228158e-01 1.04766738e+00
3.43163103e-01 -5.40596664e-01 -6.51992738e-01 -4.80174303e-01
6.36190593e-01 5.59817553e-01 -7.04402149e-01 2.57327527e-01
-4.70404476e-01 -1.05297887e+00 1.18256003e-01 3.02932769e-01
1.35769360e-02 -4.11584169e-01 -4.67579752e-01 2.26858929e-01
-8.27775896e-02 -8.13799202e-01 -9.62282479e-01 2.20715255e-01
-1.08740580e+00 -1.10569084e+00 -6.65267110e-01 -6.20024085e-01
1.11317039e+00 6.79229438e-01 7.39798486e-01 7.42172636e-03
-2.50675738e-01 4.74797457e-01 -1.64860770e-01 -4.82372463e-01
1.72610477e-01 1.36257350e-01 -5.52027524e-02 -2.88712412e-01
-7.73377763e-03 -4.13774133e-01 -5.71132481e-01 2.24403709e-01
-1.33975863e+00 -4.37982291e-01 2.01489225e-01 3.40636909e-01
9.40575600e-01 3.10081899e-01 5.82082212e-01 -1.09977460e+00
5.97606003e-01 -8.08128893e-01 -1.06860244e+00 1.60710514e-01
-8.45429659e-01 -5.73609918e-02 4.33103114e-01 -1.15282126e-02
-4.29346114e-01 -9.37196389e-02 4.88626927e-01 -9.83131677e-02
8.33674312e-01 7.36225605e-01 -4.16900277e-01 -8.90597776e-02
3.26879859e-01 3.35944712e-01 2.43642461e-02 -4.04617637e-01
4.27132726e-01 6.07786596e-01 1.15053631e-01 -6.40125751e-01
7.95978606e-01 8.37270319e-01 3.41058791e-01 -6.56978548e-01
-7.12836444e-01 -6.03795171e-01 -1.80632487e-01 1.32103994e-01
4.21435460e-02 -8.15515220e-01 -9.68793333e-01 1.65412322e-01
-8.70496035e-01 -9.98131558e-02 -5.57644248e-01 -3.86676961e-03
-8.03975046e-01 4.84802037e-01 -1.64366126e-01 -9.48078275e-01
-5.01725316e-01 -6.64950311e-01 7.27372229e-01 1.39413532e-02
2.27217041e-02 -5.22098958e-01 1.96718842e-01 5.77790260e-01
1.88249335e-01 4.22071606e-01 8.59386384e-01 -4.41073507e-01
-1.02500165e+00 -1.83280945e-01 -2.10969016e-01 1.58769056e-01
-1.77493438e-01 -3.84616613e-01 -3.14927995e-01 -8.57560515e-01
-1.11855410e-01 -2.58255273e-01 9.18061197e-01 6.15532279e-01
1.42221832e+00 -9.90061104e-01 -2.96983421e-01 7.66396999e-01
1.67688167e+00 1.81451254e-02 4.14593726e-01 2.30412230e-01
2.28792444e-01 6.30866766e-01 8.93973172e-01 1.01702595e+00
5.93044281e-01 3.87072802e-01 6.65107548e-01 3.69322121e-01
6.27387941e-01 1.78718567e-01 1.81086361e-01 3.09100151e-01
1.50348112e-01 -7.23571360e-01 -4.23338950e-01 1.11619902e+00
-2.07587290e+00 -7.90802658e-01 -2.13780299e-01 2.55218005e+00
9.99247313e-01 -1.25946522e-01 4.83692348e-01 2.43708283e-01
5.29204190e-01 2.73520261e-01 -6.49061084e-01 -7.30589509e-01
-3.31083208e-01 1.90058157e-01 1.22003520e+00 7.26168454e-01
-6.14269793e-01 4.48127389e-01 5.99501133e+00 7.57246077e-01
-5.75928390e-01 -8.17535743e-02 5.48135936e-01 -1.16022468e+00
-9.70373094e-01 4.50256132e-02 -9.50559199e-01 3.44528466e-01
6.65716410e-01 -9.18847382e-01 6.69693828e-01 7.56329060e-01
2.97448844e-01 -6.05682909e-01 -1.09485888e+00 7.79560804e-01
2.61023313e-01 -1.35091197e+00 -8.30719471e-02 5.33659756e-01
1.26192808e+00 -2.83405781e-01 1.43879756e-01 -3.21100444e-01
2.37203196e-01 -1.02130067e+00 6.15737021e-01 -3.56427953e-02
6.13482714e-01 -1.21870971e+00 4.21923071e-01 4.15625215e-01
-9.14412498e-01 -2.96809822e-01 -5.48020124e-01 5.70114888e-03
4.05719668e-01 9.62288618e-01 -6.84448302e-01 6.50081754e-01
8.36596370e-01 2.57789552e-01 1.07088983e-01 1.16302752e+00
2.85711139e-01 4.31677073e-01 -1.02149153e+00 -9.39895678e-03
1.72016367e-01 -2.29535669e-01 8.66495728e-01 1.03387070e+00
3.06433409e-01 3.93877059e-01 1.93006083e-01 6.23733103e-01
-5.72610617e-01 3.85003060e-01 -3.62511009e-01 -3.72127034e-02
5.85494101e-01 1.09334171e+00 -5.96654892e-01 -1.65828392e-01
-2.50293780e-02 4.67611462e-01 1.84214473e-01 2.09753081e-01
-5.84992290e-01 -3.46595585e-01 7.04684794e-01 5.96872330e-01
7.14504778e-01 -3.32973570e-01 -9.15596545e-01 -6.84570074e-01
4.78089869e-01 -6.25448465e-01 8.48577380e-01 -1.79464521e-03
-9.46231008e-01 -1.70069695e-01 4.22398716e-01 -8.57946634e-01
2.30812818e-01 -1.37481138e-01 -2.34500721e-01 7.37829566e-01
-1.41921031e+00 -6.34239972e-01 -9.72861201e-02 4.61048752e-01
1.18552402e-01 4.73531038e-01 8.93262923e-02 1.27625987e-01
-3.86165112e-01 7.72217333e-01 4.76497144e-01 -6.83025241e-01
3.00782502e-01 -1.25016630e+00 -4.09087390e-01 8.35042059e-01
-4.08089161e-01 6.05361700e-01 9.56151009e-01 -5.29376388e-01
-1.83307970e+00 -9.75484014e-01 9.39082623e-01 1.67660303e-02
4.30653036e-01 -3.91070694e-01 -5.51196635e-01 7.78495669e-01
1.59566164e-01 -1.57690674e-01 8.07643890e-01 -3.28834206e-02
-1.54613107e-01 -4.48075384e-01 -1.70558679e+00 2.99976885e-01
1.29291856e+00 2.10855618e-01 -4.52950299e-02 4.45770323e-01
8.47421885e-01 -5.16290307e-01 -9.33554649e-01 5.18481314e-01
4.22199786e-01 -5.02363980e-01 7.75644302e-01 -4.91135091e-01
3.46680731e-01 -2.95577347e-01 -6.06206536e-01 -1.01999247e+00
-7.36317709e-02 -9.74828482e-01 -5.09346843e-01 8.34996641e-01
6.25193775e-01 -5.48622191e-01 9.90878046e-01 6.46792233e-01
8.02876428e-02 -1.12759483e+00 -1.23401463e+00 -7.43076801e-01
2.75283217e-01 -2.45641991e-01 6.07977509e-01 7.05823481e-01
2.40934014e-01 -8.83864984e-02 -5.13257444e-01 3.26836109e-01
9.60785031e-01 6.27836883e-01 9.50813830e-01 -8.88185620e-01
-4.69643563e-01 -2.75448710e-01 5.14818132e-02 -1.28149462e+00
-3.21320206e-01 -1.12880087e+00 -2.09560752e-01 -1.61563218e+00
7.19582975e-01 -5.84708810e-01 -1.54683888e-01 3.33517462e-01
3.97129431e-02 7.57433474e-02 5.30733824e-01 -3.94292511e-02
-9.35639143e-01 5.52878022e-01 1.30637562e+00 -2.43918568e-01
-4.75639522e-01 6.29616082e-02 -1.45794988e+00 2.65739739e-01
5.99382699e-01 -7.01578736e-01 -3.71553928e-01 -4.93214011e-01
6.54024303e-01 3.79969776e-01 -3.18201408e-02 -2.70274818e-01
3.44578147e-01 -4.54562366e-01 -2.45994538e-01 -9.74397659e-01
8.95055458e-02 -8.54784131e-01 1.04929075e-01 4.62463439e-01
-2.70195931e-01 1.08037613e-01 2.35327482e-01 5.82709253e-01
2.92955935e-02 -4.31340843e-01 7.75049567e-01 -1.65186021e-02
8.23680013e-02 5.68701923e-01 -2.02542860e-02 3.54398072e-01
1.55486226e+00 -4.27207500e-01 -4.68107224e-01 -5.16198456e-01
-4.71416205e-01 1.05131018e+00 5.40974557e-01 -1.87268645e-01
4.41072494e-01 -8.84940386e-01 -8.99243772e-01 -4.27274257e-01
8.89968947e-02 6.29512191e-01 4.43513006e-01 1.16467392e+00
-5.43500125e-01 5.94902277e-01 3.13037038e-01 -3.31151038e-01
-1.28039026e+00 6.65899873e-01 5.93348257e-02 -3.90424997e-01
-2.33935609e-01 9.37405050e-01 8.43314976e-02 -2.42897987e-01
2.97102511e-01 -1.94157973e-01 5.28810680e-01 3.65562320e-01
4.91649538e-01 9.25266862e-01 -7.33480230e-02 -2.51866557e-04
-3.63322735e-01 2.26857513e-01 -3.63688409e-01 -2.10565627e-01
1.66554785e+00 -3.57544690e-01 -6.07113302e-01 -1.20123792e-02
1.17319214e+00 4.12608147e-01 -8.78831506e-01 -3.30822974e-01
-2.06788525e-01 -6.67701602e-01 -1.51150197e-01 -5.82882702e-01
-1.23041821e+00 -1.23560978e-02 -7.24624842e-02 4.75812882e-01
1.37331307e+00 3.87896180e-01 1.07591939e+00 3.65513206e-01
7.37238050e-01 -1.28147209e+00 -3.80898982e-01 1.40188456e-01
8.97653043e-01 -8.00829649e-01 6.16812408e-01 -6.80553317e-01
-1.71034545e-01 8.02949607e-01 3.52046847e-01 -2.47870475e-01
6.50365174e-01 3.96113068e-01 -5.27141452e-01 -3.15656513e-01
-9.09929276e-01 -5.04161306e-02 -1.41982675e-01 -1.64500736e-02
-7.63005316e-02 2.18580142e-01 -1.19426155e+00 6.77903652e-01
-3.44048768e-01 -1.12435445e-01 8.77743542e-01 1.23910415e+00
-9.25594568e-01 -1.01182234e+00 -6.64819777e-01 7.97775030e-01
-5.09714544e-01 -2.14381777e-02 -4.60404158e-01 5.57434440e-01
4.95031364e-02 1.06311476e+00 -2.25161593e-02 6.30761608e-02
2.41644815e-01 -4.12960857e-01 6.43868208e-01 -5.42059600e-01
-4.69840437e-01 3.10943145e-02 5.73561043e-02 -4.40155387e-01
-2.09382668e-01 -9.25252318e-01 -1.33241796e+00 -6.93443060e-01
-2.44445354e-01 2.33644873e-01 4.72980469e-01 6.84959054e-01
2.35843524e-01 1.95033588e-02 9.33383763e-01 -1.51940987e-01
-9.02347326e-01 -6.72687232e-01 -1.17634499e+00 -3.11696585e-02
2.60014832e-01 -2.05273107e-01 -5.03821015e-01 -4.01193976e-01]
|
[6.587095737457275, 4.895382881164551]
|
e8a50cbc-6332-45ac-9944-e3693ba2a001
|
adversarial-generation-of-training-examples
|
1707.03124
| null |
http://arxiv.org/abs/1707.03124v3
|
http://arxiv.org/pdf/1707.03124v3.pdf
|
Adversarial Generation of Training Examples: Applications to Moving Vehicle License Plate Recognition
|
Generative Adversarial Networks (GAN) have attracted much research attention
recently, leading to impressive results for natural image generation. However,
to date little success was observed in using GAN generated images for improving
classification tasks. Here we attempt to explore, in the context of car license
plate recognition, whether it is possible to generate synthetic training data
using GAN to improve recognition accuracy. With a carefully-designed pipeline,
we show that the answer is affirmative. First, a large-scale image set is
generated using the generator of GAN, without manual annotation. Then, these
images are fed to a deep convolutional neural network (DCNN) followed by a
bidirectional recurrent neural network (BRNN) with long short-term memory
(LSTM), which performs the feature learning and sequence labelling. Finally,
the pre-trained model is fine-tuned on real images. Our experimental results on
a few data sets demonstrate the effectiveness of using GAN images: an
improvement of 7.5% over a strong baseline with moderate-sized real data being
available. We show that the proposed framework achieves competitive recognition
accuracy on challenging test datasets. We also leverage the depthwise separate
convolution to construct a lightweight convolutional RNN, which is about half
size and 2x faster on CPU. Combining this framework and the proposed pipeline,
we make progress in performing accurate recognition on mobile and embedded
devices.
|
['Chunhua Shen', 'Zhipeng Man', 'Xinlong Wang', 'Mingyu You']
|
2017-07-11
| null | null | null | null |
['license-plate-recognition']
|
['computer-vision']
|
[ 8.08886468e-01 1.46463886e-01 1.29604653e-01 -2.38497853e-01
-1.12758660e+00 -5.46082437e-01 7.73638546e-01 -8.54896724e-01
-3.48527968e-01 8.41351807e-01 1.69391483e-02 -5.35672069e-01
7.25265026e-01 -8.98810148e-01 -1.07913601e+00 -8.76306415e-01
2.81980336e-01 1.68346226e-01 -1.99608691e-02 5.95950112e-02
1.71634749e-01 4.43374962e-01 -1.44229829e+00 4.01319265e-01
5.92192829e-01 1.01065600e+00 3.25614177e-02 9.26083446e-01
2.05762178e-01 1.11280107e+00 -9.27770078e-01 -6.98588192e-01
5.07262290e-01 -6.23766780e-01 -6.66987062e-01 2.76720107e-01
3.19298506e-01 -4.69587743e-01 -4.27747458e-01 7.62752533e-01
5.44110239e-01 -6.42855987e-02 4.69176799e-01 -1.01517332e+00
-8.61366928e-01 3.10762137e-01 -2.72388011e-01 -6.77326545e-02
4.83874381e-02 6.17100835e-01 5.38149655e-01 -8.98018479e-01
4.26922828e-01 7.10584164e-01 6.67978466e-01 9.23992157e-01
-8.68421555e-01 -7.12148368e-01 -3.57124418e-01 -8.78587738e-02
-1.27108192e+00 -4.76909965e-01 6.60255313e-01 -1.45068541e-01
1.02340114e+00 1.82261720e-01 5.09905994e-01 1.62460172e+00
-6.53392228e-04 7.72555292e-01 1.27235126e+00 -4.91248250e-01
2.24126562e-01 4.24067751e-02 -2.81942785e-01 5.66891909e-01
-5.05969338e-02 1.60930350e-01 -1.80100352e-01 2.01655567e-01
8.51046383e-01 1.48457438e-01 -2.46345431e-01 1.35341987e-01
-1.18022060e+00 8.46418142e-01 5.22984862e-01 1.53296903e-01
-5.30216992e-01 4.71808434e-01 4.30649966e-01 2.29478091e-01
4.20424640e-01 3.59057188e-01 -3.94211113e-02 -3.30812126e-01
-9.06924725e-01 4.58229110e-02 6.31054163e-01 9.78688419e-01
5.55901587e-01 5.66897690e-01 -2.30740339e-01 9.20275092e-01
-5.75061247e-04 5.76056182e-01 8.26229453e-01 -6.61500454e-01
5.30784726e-01 2.62947202e-01 -3.65271010e-02 -7.10997522e-01
1.31222531e-01 -5.64947188e-01 -1.13540852e+00 5.06968200e-01
3.36903304e-01 -3.52351636e-01 -1.26773441e+00 1.33549154e+00
-1.28919214e-01 5.99032402e-01 4.68449473e-01 8.03921282e-01
4.92523193e-01 7.92363346e-01 -2.49392726e-02 2.27696046e-01
1.25929785e+00 -1.35248327e+00 -2.52325505e-01 -3.37797046e-01
6.22843504e-01 -7.26046979e-01 1.03452647e+00 3.35379690e-01
-1.04192436e+00 -7.41697013e-01 -1.31474423e+00 2.17645705e-01
-3.03042561e-01 2.98569858e-01 5.53644657e-01 9.26276326e-01
-1.13616455e+00 3.25002551e-01 -7.64765680e-01 -1.52750790e-01
7.05796719e-01 3.63076895e-01 -3.16872478e-01 -2.84915656e-01
-8.85473609e-01 6.63006723e-01 2.27633834e-01 3.81520212e-01
-1.21496677e+00 -4.14736360e-01 -7.22103596e-01 -1.20348863e-01
-4.99779098e-02 -7.85358548e-01 1.16251791e+00 -1.43299162e+00
-2.02491093e+00 8.08239400e-01 1.69732291e-02 -8.87008190e-01
5.54860771e-01 -8.73802155e-02 -5.00680387e-01 3.82926464e-02
-2.50367165e-01 8.60774100e-01 1.16711426e+00 -1.16966736e+00
-5.11768401e-01 -9.30265617e-03 -1.66723765e-02 1.89294554e-02
-2.22322077e-01 3.25064000e-04 -4.74839926e-01 -8.85919571e-01
-4.34980452e-01 -1.26262212e+00 -2.50904799e-01 -5.20771384e-01
-4.98626709e-01 1.38518423e-01 8.21924210e-01 -8.00950050e-01
6.66880131e-01 -2.17533851e+00 -2.53423572e-01 1.95778564e-01
-1.47252366e-01 8.03281426e-01 -3.34068865e-01 1.79312006e-01
-2.07142979e-01 1.79823428e-01 -4.04755682e-01 -4.77687597e-01
-9.32536870e-02 1.35663196e-01 -7.38796771e-01 1.84881672e-01
6.07138574e-01 1.45581484e+00 -4.34980392e-01 1.87286332e-01
1.32109359e-01 6.84033811e-01 -3.54865432e-01 4.59241182e-01
-7.56179690e-02 4.38401282e-01 -1.42644152e-01 6.04332328e-01
6.04731858e-01 -2.12522745e-01 -1.02670163e-01 1.17526673e-01
2.49145225e-01 1.77689627e-01 -6.37408435e-01 1.34535408e+00
-5.96163690e-01 8.57805133e-01 -4.05265629e-01 -1.08367252e+00
1.11918330e+00 2.51736671e-01 -1.05159111e-01 -8.50085557e-01
6.81688339e-02 2.40397960e-01 -1.04084678e-01 -1.66783154e-01
2.83907980e-01 -1.09174609e-01 -1.48621529e-01 4.57845658e-01
1.51882870e-02 -8.16397369e-02 -1.91979289e-01 -1.06476851e-01
1.36456335e+00 1.92154154e-01 -1.27059430e-01 3.12973499e-01
5.80348730e-01 -2.29706932e-02 1.02513038e-01 7.17852175e-01
1.21664591e-01 9.90855813e-01 2.84215033e-01 -5.73205113e-01
-1.44595647e+00 -7.70207107e-01 3.46185207e-01 7.17983365e-01
-2.76096582e-01 -4.34416831e-02 -1.10317266e+00 -8.49581003e-01
-5.40144503e-01 6.99651837e-01 -6.76650226e-01 -1.63184777e-01
-8.29672098e-01 -8.29256415e-01 1.08190286e+00 7.51788616e-01
9.73149896e-01 -1.42239666e+00 -4.59952086e-01 6.84920326e-02
2.81595260e-01 -1.20559597e+00 -3.06723744e-01 -6.59986213e-03
-7.48327315e-01 -6.83733821e-01 -9.56401587e-01 -9.63465095e-01
7.58304954e-01 6.00678213e-02 1.01848006e+00 1.72903284e-01
-1.42483905e-01 1.24139465e-01 -3.07153940e-01 -1.90249786e-01
-7.37619162e-01 2.48645335e-01 -3.40705484e-01 1.50015622e-01
1.71537533e-01 -6.18394971e-01 -6.56222403e-01 3.07781398e-01
-1.05194449e+00 3.77766848e-01 1.01081848e+00 1.06292820e+00
3.90106052e-01 -2.20477164e-01 6.77672148e-01 -9.25586820e-01
5.21198213e-01 -3.21388692e-01 -6.83389425e-01 7.38330651e-04
-3.73467386e-01 -4.25997339e-02 8.75658989e-01 -3.61881822e-01
-1.05250096e+00 1.92865714e-01 -7.11749673e-01 -4.26564097e-01
-3.93804938e-01 2.77855694e-01 -1.71478912e-01 -1.43176436e-01
5.41938961e-01 6.10806763e-01 1.00149155e-01 -7.45038390e-02
3.49146605e-01 9.17908907e-01 7.91512847e-01 -1.85520649e-01
8.68366659e-01 3.39856297e-01 -2.89980769e-01 -6.19938254e-01
-6.10101342e-01 1.44762695e-01 -2.73163199e-01 2.64155902e-02
9.17861164e-01 -1.03533483e+00 -6.68468952e-01 8.52693021e-01
-9.83229160e-01 -7.16987371e-01 -1.93838760e-01 2.41549492e-01
-5.72063267e-01 -6.65524006e-02 -7.92893767e-01 -6.58296824e-01
-6.48901463e-01 -1.12681580e+00 1.09993279e+00 1.65428981e-01
1.54905409e-01 -8.25197101e-01 -1.21087134e-01 6.79767668e-01
7.49975324e-01 3.43876600e-01 4.29669052e-01 -7.32479751e-01
-9.48761225e-01 -4.25808191e-01 -2.03863025e-01 8.48012269e-01
-6.36433158e-03 -1.94600195e-01 -1.20418787e+00 -2.06774369e-01
2.20787339e-02 -5.10488689e-01 9.86776710e-01 3.50050144e-02
1.27928412e+00 -4.28636014e-01 -8.87392834e-02 7.66668916e-01
1.26942039e+00 3.01201999e-01 1.29014754e+00 2.53853500e-01
8.61889899e-01 8.23135227e-02 2.24946141e-01 -2.03456804e-02
1.19049631e-01 6.06602848e-01 3.96077067e-01 -3.38250935e-01
-3.93137842e-01 -3.85998785e-01 5.90261459e-01 7.49448657e-01
-2.19954103e-01 -3.52175534e-01 -9.10188973e-01 3.67008924e-01
-1.44220197e+00 -1.09655190e+00 1.77273497e-01 2.08047462e+00
6.84571087e-01 3.31084967e-01 6.88597709e-02 8.70117024e-02
5.76585650e-01 -2.82749012e-02 -4.95751143e-01 -4.32813168e-01
-1.01679288e-01 5.06431699e-01 6.13712192e-01 3.73412520e-01
-1.08166373e+00 1.09502721e+00 6.28008366e+00 8.98574054e-01
-1.60037327e+00 8.46407637e-02 1.15539801e+00 2.46250376e-01
-2.11491838e-01 -1.87170729e-01 -7.05758989e-01 6.08472049e-01
1.35722828e+00 2.04225272e-01 5.27698755e-01 9.36750114e-01
-1.49231493e-01 3.11026871e-01 -8.13895524e-01 9.53559756e-01
3.48169059e-01 -1.55616283e+00 1.12486139e-01 1.48769349e-01
8.41612637e-01 2.68814802e-01 2.49902189e-01 5.39485574e-01
3.67826521e-01 -1.50260162e+00 4.67876256e-01 5.51691592e-01
9.88243759e-01 -1.07871592e+00 9.94393051e-01 4.28124845e-01
-7.40957916e-01 1.25035480e-01 -2.38584638e-01 -6.92281872e-02
3.24673615e-02 4.11545694e-01 -1.31985402e+00 3.27257484e-01
3.27210248e-01 5.68037510e-01 -6.75985992e-01 6.96183205e-01
-4.29656923e-01 1.05666029e+00 3.49634537e-03 8.38385597e-02
3.50602567e-01 -9.81871635e-02 5.12136072e-02 1.32399845e+00
4.66839284e-01 -8.57090876e-02 -3.62924695e-01 8.41543674e-01
-4.45772976e-01 -3.31115097e-01 -7.81920612e-01 -1.81820348e-01
1.18795536e-01 1.33454895e+00 -6.90592051e-01 -5.06570935e-01
-4.21522945e-01 1.53477061e+00 2.90027857e-01 4.56270486e-01
-1.08820200e+00 -3.66390795e-01 4.28270221e-01 -1.86312556e-01
6.71514809e-01 -1.37736648e-01 -2.44559854e-01 -1.11382723e+00
8.27252567e-02 -1.10669732e+00 2.24519335e-03 -8.40116382e-01
-1.11628616e+00 1.02837181e+00 -6.29417658e-01 -1.32586014e+00
-7.69586205e-01 -6.51030004e-01 -7.36430883e-01 1.02997971e+00
-1.29038751e+00 -1.36300111e+00 -4.02551264e-01 4.29473877e-01
7.12758720e-01 -4.88107443e-01 9.85017240e-01 3.03949147e-01
-5.95930398e-01 8.64747643e-01 -4.20002546e-03 5.04496634e-01
3.49859983e-01 -9.46310401e-01 1.03809309e+00 1.03982127e+00
2.09202096e-01 4.47428495e-01 3.05622280e-01 -4.87552971e-01
-1.50262606e+00 -1.52669251e+00 5.18275797e-01 -4.87471133e-01
3.41453582e-01 -6.84664249e-01 -8.58263552e-01 7.56992877e-01
4.90781546e-01 2.74316162e-01 5.90866089e-01 -6.29728138e-01
-2.79403210e-01 -7.58293271e-02 -1.14359486e+00 5.40216982e-01
9.36973035e-01 -5.57260156e-01 -3.24788541e-01 2.50212222e-01
7.95541048e-01 -5.14415205e-01 -6.61644101e-01 2.75339127e-01
5.09621263e-01 -8.59574676e-01 8.63583505e-01 -4.79519010e-01
7.68011570e-01 -4.09322500e-01 -1.00960031e-01 -1.33580017e+00
6.44354150e-02 -6.23631775e-01 2.67620701e-02 1.31611025e+00
5.28749704e-01 -7.25068986e-01 1.00248623e+00 5.25103867e-01
-3.41378957e-01 -8.69343042e-01 -7.14164555e-01 -6.64916635e-01
1.02181742e-02 -5.47380805e-01 6.19256198e-01 5.60717046e-01
-4.93770599e-01 4.91722822e-01 -7.23719537e-01 6.77545220e-02
2.49121830e-01 2.61802394e-02 1.14754009e+00 -4.61234152e-01
-4.79315758e-01 -1.29613966e-01 -6.21532142e-01 -1.17977989e+00
2.69184232e-01 -7.25714684e-01 8.53230581e-02 -1.19002974e+00
2.14819070e-02 -3.65813911e-01 -8.27203616e-02 5.05732954e-01
-1.55060366e-01 1.01581895e+00 2.63098955e-01 2.35140651e-01
-4.66882944e-01 5.44026971e-01 1.07462323e+00 -2.73420215e-01
2.42490485e-01 1.16859220e-01 -6.30743384e-01 6.17202580e-01
8.44350040e-01 -2.50543743e-01 -2.89712906e-01 -4.08676177e-01
-5.77305593e-02 -8.54379013e-02 5.47884941e-01 -1.19405556e+00
1.06430814e-01 3.46559346e-01 6.97720170e-01 -2.44442433e-01
3.62168431e-01 -6.86454952e-01 3.18548173e-01 4.44618642e-01
-2.63307005e-01 4.46976200e-02 2.30278358e-01 3.63716215e-01
-3.45186800e-01 -1.05669454e-01 6.30399227e-01 1.31189283e-02
-6.18419528e-01 2.09391370e-01 -3.38318139e-01 -9.36418697e-02
9.98590350e-01 -2.28884786e-01 -2.17800781e-01 -5.53329587e-01
-4.14849907e-01 -3.88945341e-01 4.96851653e-01 4.89433050e-01
6.62297308e-01 -1.48399067e+00 -8.08944702e-01 5.24402797e-01
-8.29517841e-02 -2.55995598e-02 9.18595493e-02 4.17061359e-01
-6.90742493e-01 4.62133944e-01 -2.57235438e-01 -5.83719850e-01
-9.43106949e-01 5.28051853e-01 3.83671075e-01 -3.44129175e-01
-5.31426668e-01 7.77100801e-01 1.09198987e-01 -3.68707418e-01
-3.05271037e-02 -9.93835628e-02 1.02277249e-01 -3.92618030e-01
6.85138345e-01 4.71566524e-03 2.54961967e-01 -5.66658318e-01
-1.08119130e-01 2.49802426e-01 -5.74629828e-02 -2.47988194e-01
1.44508517e+00 3.16526592e-01 1.52546301e-01 3.59473005e-02
1.29820728e+00 -7.01154247e-02 -1.52212858e+00 7.02670664e-02
-2.06329480e-01 -3.85314941e-01 -2.78920054e-01 -6.94645226e-01
-1.30299115e+00 7.54516125e-01 6.16353750e-01 7.99107477e-02
1.29451096e+00 -2.91975737e-01 9.93970215e-01 4.47805732e-01
3.85472357e-01 -5.80933452e-01 1.53386429e-01 4.87418115e-01
8.61297727e-01 -1.11412501e+00 -6.06211364e-01 -1.97654501e-01
-8.21754932e-01 9.96276200e-01 6.25176907e-01 -5.38625836e-01
2.46280320e-02 5.11420488e-01 1.55776069e-01 1.90850496e-01
-6.31009519e-01 1.34427220e-01 1.43127993e-01 6.03567123e-01
3.54247808e-01 1.46211889e-02 1.83664739e-01 3.88784975e-01
-3.93587232e-01 1.79041043e-01 5.52390695e-01 7.05838263e-01
8.77069756e-02 -1.18406725e+00 -2.30663896e-01 4.44989830e-01
-5.68962991e-01 -3.36580664e-01 -2.54144728e-01 7.03135133e-01
-3.35508287e-02 6.25740826e-01 2.03011319e-01 -4.45903718e-01
1.71497002e-01 1.92095026e-01 2.78952599e-01 -4.59618837e-01
-6.60001993e-01 -2.52644271e-01 8.26706365e-02 -5.16812742e-01
-2.79824585e-01 -4.39306796e-01 -1.00894773e+00 -1.78757221e-01
-1.52796239e-01 -2.13736128e-02 7.66706705e-01 9.85271454e-01
5.42684078e-01 5.98230481e-01 7.19680667e-01 -9.07088280e-01
-4.00878012e-01 -1.01358485e+00 -5.12744188e-02 4.51384366e-01
3.08660984e-01 -2.11596210e-02 -2.82726884e-01 3.96490395e-01]
|
[11.648541450500488, -0.507445216178894]
|
1b2d7f5a-f864-4895-bbb3-a1d53bffe1ac
|
adversarial-autoencoders
|
1511.05644
| null |
http://arxiv.org/abs/1511.05644v2
|
http://arxiv.org/pdf/1511.05644v2.pdf
|
Adversarial Autoencoders
|
In this paper, we propose the "adversarial autoencoder" (AAE), which is a
probabilistic autoencoder that uses the recently proposed generative
adversarial networks (GAN) to perform variational inference by matching the
aggregated posterior of the hidden code vector of the autoencoder with an
arbitrary prior distribution. Matching the aggregated posterior to the prior
ensures that generating from any part of prior space results in meaningful
samples. As a result, the decoder of the adversarial autoencoder learns a deep
generative model that maps the imposed prior to the data distribution. We show
how the adversarial autoencoder can be used in applications such as
semi-supervised classification, disentangling style and content of images,
unsupervised clustering, dimensionality reduction and data visualization. We
performed experiments on MNIST, Street View House Numbers and Toronto Face
datasets and show that adversarial autoencoders achieve competitive results in
generative modeling and semi-supervised classification tasks.
|
['Brendan Frey', 'Navdeep Jaitly', 'Jonathon Shlens', 'Alireza Makhzani', 'Ian Goodfellow']
|
2015-11-18
| null | null | null | null |
['unsupervised-image-classification', 'unsupervised-mnist']
|
['computer-vision', 'methodology']
|
[ 5.15548959e-02 5.64532459e-01 5.85874438e-01 -4.22627032e-01
-4.28123683e-01 -5.87726772e-01 8.87894452e-01 -7.06902087e-01
1.11923642e-01 6.89970791e-01 3.80020469e-01 1.11732192e-01
3.52485068e-02 -1.03349125e+00 -1.06711912e+00 -9.47710872e-01
1.79079220e-01 1.06741929e+00 -3.24188292e-01 1.57500070e-03
-3.26344788e-01 4.58479673e-01 -1.33863401e+00 3.85965139e-01
5.51769137e-01 5.59663534e-01 -4.64904830e-02 1.02207708e+00
1.05359890e-01 1.07462597e+00 -9.71867025e-01 -6.61349356e-01
2.32566908e-01 -7.31633246e-01 -6.12344980e-01 2.65060186e-01
1.53310478e-01 -4.42803115e-01 -6.86057806e-01 1.34326673e+00
2.53739983e-01 1.11288734e-01 1.34425664e+00 -1.54253924e+00
-1.34015739e+00 6.47161722e-01 -8.83547142e-02 -2.20952004e-01
-2.46606376e-02 -6.44662082e-02 7.80482888e-01 -6.05971575e-01
6.34198546e-01 1.53676403e+00 5.19398630e-01 7.04057336e-01
-1.68570125e+00 -5.99103570e-01 -4.93646085e-01 -1.18188463e-01
-1.54934883e+00 -2.32007354e-01 1.00574720e+00 -6.54999971e-01
7.29491889e-01 6.95400238e-02 6.23273134e-01 1.64204323e+00
4.23082411e-01 5.59512556e-01 1.02572298e+00 -2.67038852e-01
6.51268482e-01 3.32110554e-01 -6.53961837e-01 6.69823408e-01
-2.29973868e-01 2.63337076e-01 -1.84851974e-01 -3.06583315e-01
1.04326475e+00 2.28794143e-01 1.21334255e-01 -5.20430207e-01
-7.68305004e-01 1.39225733e+00 4.35494930e-01 1.21875763e-01
-4.35072392e-01 5.72089434e-01 -5.22210114e-02 2.64187396e-01
4.94816780e-01 2.85576016e-01 5.01284096e-03 1.94999069e-01
-8.54940057e-01 1.60387754e-01 8.30014825e-01 1.10310888e+00
7.52790332e-01 6.61259055e-01 -1.08760871e-01 7.16562867e-01
7.03098655e-01 7.58934319e-01 6.08304977e-01 -1.24016392e+00
-1.87461779e-01 3.13481957e-01 -3.40291560e-01 -7.39849865e-01
2.65244812e-01 -2.08954662e-01 -1.04397929e+00 6.89343512e-01
9.90194008e-02 -3.72960299e-01 -1.21164966e+00 1.67935634e+00
1.67213932e-01 3.78441423e-01 3.30456555e-01 5.09595752e-01
6.93867564e-01 7.36994207e-01 -5.87345324e-02 2.79228091e-01
1.13724339e+00 -3.04055333e-01 -8.25335383e-01 -1.10390440e-01
-1.91098899e-01 -6.81003928e-01 5.92266679e-01 2.08605781e-01
-9.61241722e-01 -6.41070247e-01 -1.16148686e+00 -5.65756299e-02
-5.85107803e-01 9.16581824e-02 3.68931264e-01 8.18483829e-01
-1.12022305e+00 6.43479705e-01 -1.24137890e+00 -8.66095796e-02
7.36344278e-01 4.37500238e-01 -3.04055899e-01 1.52510673e-01
-9.38523293e-01 5.24728179e-01 5.17582238e-01 -3.77173781e-01
-1.60939443e+00 -6.74397945e-01 -9.22404945e-01 1.36268690e-01
-3.59815687e-01 -8.77384543e-01 8.99455905e-01 -1.20909858e+00
-1.70307267e+00 8.59011948e-01 1.65013909e-01 -7.72989869e-01
3.29923898e-01 -7.00942427e-02 -3.94685239e-01 2.39255410e-02
-1.79461613e-01 8.88832033e-01 1.67347193e+00 -1.37888443e+00
-6.25173897e-02 -3.69574249e-01 -2.71304667e-01 -1.07878551e-01
-1.10439122e-01 -4.26584274e-01 -2.19355375e-02 -9.92328048e-01
-2.26871088e-01 -1.00155652e+00 2.96636950e-02 -3.16419750e-01
-7.05555618e-01 8.55535343e-02 1.38448131e+00 -6.15624189e-01
5.71743846e-01 -2.32750487e+00 5.55878401e-01 3.68724674e-01
4.01490778e-01 -1.33410826e-01 1.55896917e-01 4.07855511e-01
-4.09191012e-01 1.68549158e-02 -5.06435633e-01 -6.58062398e-01
3.65289718e-01 7.98528075e-01 -5.89550793e-01 5.83800077e-01
3.75141948e-01 9.53501642e-01 -6.21450424e-01 -3.34788918e-01
3.80646646e-01 1.13354552e+00 -8.41934860e-01 6.85531795e-01
-3.51830840e-01 5.13131440e-01 -4.91217598e-02 2.79636174e-01
5.82120180e-01 -3.20660807e-02 5.77553846e-02 -4.29586321e-02
6.44749105e-01 -2.69191712e-01 -9.60287333e-01 1.63137662e+00
-1.42634064e-01 8.72346759e-01 -3.08911297e-02 -9.24398303e-01
9.11046565e-01 3.80368918e-01 2.05359250e-01 6.19183555e-02
3.03658724e-01 -3.25553298e-01 -5.18838577e-02 -2.56429225e-01
1.56705841e-01 -3.38426292e-01 1.43176140e-02 6.00157142e-01
8.73673260e-01 -4.56626981e-01 -3.40628922e-01 3.58025312e-01
7.76514888e-01 1.99050099e-01 1.24006771e-01 -3.77637208e-01
9.18230042e-02 -3.53119850e-01 1.59825742e-01 5.86671114e-01
9.84390154e-02 9.06538308e-01 4.84227628e-01 -2.75671065e-01
-1.42367005e+00 -1.72183967e+00 4.22204435e-02 8.02695334e-01
-6.10552251e-01 -1.86996222e-01 -1.29471624e+00 -7.84830272e-01
-6.40976429e-02 1.08656311e+00 -1.18155086e+00 -4.69610423e-01
-1.85271040e-01 -6.35909915e-01 6.15772605e-01 6.61869943e-01
3.76974255e-01 -9.85962093e-01 -1.92303043e-02 4.60672528e-02
1.89775974e-01 -9.71341908e-01 -2.27530450e-01 2.86719531e-01
-6.45627797e-01 -8.48459482e-01 -6.05067909e-01 -6.70806587e-01
8.16974998e-01 -6.49628103e-01 1.01489556e+00 -4.91448164e-01
-2.38050282e-01 5.40528119e-01 -2.18150452e-01 -5.26801109e-01
-1.14972150e+00 -1.84115365e-01 1.23432837e-02 2.97569841e-01
4.83960599e-01 -1.11472189e+00 -2.40416571e-01 -3.98590527e-02
-1.21566391e+00 -1.29866496e-01 2.16591477e-01 1.00957060e+00
6.38391972e-01 2.84796089e-01 3.01126629e-01 -1.04355192e+00
5.07622182e-01 -7.15426207e-01 -6.16097867e-01 -4.67387587e-02
-2.97838837e-01 4.81364667e-01 8.14189851e-01 -3.98096174e-01
-1.04233551e+00 1.50317699e-01 -4.02804792e-01 -1.17293751e+00
-4.90939826e-01 2.26477869e-02 -3.90317500e-01 3.95012587e-01
7.81284213e-01 3.63937974e-01 2.33399644e-01 -3.21712643e-01
6.72830343e-01 5.37575066e-01 9.52901840e-01 -3.82805526e-01
1.01287091e+00 8.71444166e-01 -1.29216900e-02 -6.33414268e-01
-3.50898743e-01 3.29784185e-01 -6.56314969e-01 1.08989254e-01
1.48320603e+00 -8.52196932e-01 -4.54294741e-01 2.55214602e-01
-1.05729163e+00 -2.62615055e-01 -8.91908348e-01 2.55039871e-01
-1.00601530e+00 -9.01379138e-02 -4.45561767e-01 -7.47062922e-01
-1.15872249e-01 -1.18865061e+00 1.06738222e+00 1.91830933e-01
-3.08987677e-01 -1.45835352e+00 3.97156090e-01 -1.10963769e-02
2.23364413e-01 7.28855789e-01 7.99406648e-01 -1.05640113e+00
-5.06666899e-01 -3.28344613e-01 4.01017129e-01 6.98537111e-01
9.42735374e-02 2.68863708e-01 -1.38641870e+00 -2.16682076e-01
1.68947041e-01 -2.16183871e-01 7.21403241e-01 4.39932406e-01
1.34512091e+00 -6.95409536e-01 -9.42529142e-02 1.06043851e+00
1.33115005e+00 3.09105456e-01 1.08499157e+00 -2.67186403e-01
9.00522172e-01 4.67399836e-01 -3.68343085e-01 5.06832302e-01
2.44034622e-02 2.63426244e-01 6.79285526e-01 6.48806840e-02
-1.41561091e-01 -6.89353406e-01 3.96929562e-01 5.47805250e-01
-2.75185239e-02 -3.18249047e-01 -6.22051775e-01 4.09637541e-01
-1.55743587e+00 -1.19676030e+00 3.71570677e-01 1.74663699e+00
6.42041445e-01 -1.81088224e-01 -1.00227214e-01 3.77027020e-02
5.85557640e-01 4.21300530e-02 -5.53054392e-01 -5.60052156e-01
-9.35066193e-02 6.43280327e-01 2.67791241e-01 6.08584285e-01
-1.13493633e+00 8.76952231e-01 6.84514952e+00 9.50078547e-01
-6.26073778e-01 3.47182035e-01 5.68901062e-01 -4.64753248e-02
-4.20344830e-01 -2.10055500e-01 -4.91380274e-01 7.17295587e-01
1.11229181e+00 3.91053408e-03 9.27489519e-01 1.03656471e+00
-3.98963004e-01 4.65469033e-01 -1.48566651e+00 9.98272717e-01
2.22464517e-01 -1.34973776e+00 2.71568894e-01 4.31810886e-01
1.08990204e+00 1.54178264e-03 5.65110445e-01 2.85292387e-01
1.30960917e+00 -1.57926297e+00 5.95972478e-01 8.37823987e-01
7.58839071e-01 -1.20447373e+00 5.69898486e-01 2.45150536e-01
-5.69673896e-01 1.90062478e-01 -4.08345729e-01 3.07897300e-01
-2.35223070e-01 3.42679471e-01 -9.69785094e-01 3.30132619e-02
7.06781745e-01 5.15453160e-01 -3.38990062e-01 5.92593960e-02
-3.52711439e-01 6.93404436e-01 -2.03326315e-01 3.39130133e-01
3.90587002e-02 -4.06271607e-01 7.36256540e-01 1.02274561e+00
2.19989434e-01 -8.45090952e-03 -1.60442919e-01 1.64838994e+00
-3.89879018e-01 -5.34403861e-01 -9.85945702e-01 -3.72552484e-01
4.26311523e-01 8.75520885e-01 -4.67473149e-01 -3.78649533e-01
1.68717373e-02 1.35845888e+00 2.62448490e-01 6.68296218e-01
-8.34474742e-01 -2.33556643e-01 9.88840997e-01 -2.90330470e-01
6.82318866e-01 1.24914832e-02 -2.10293218e-01 -1.03588927e+00
-6.72074676e-01 -8.23734879e-01 2.86113709e-01 -1.07931101e+00
-1.41491604e+00 7.25698054e-01 4.23492678e-02 -1.07428014e+00
-9.95461583e-01 -6.13575876e-01 -6.83220327e-01 8.95561278e-01
-5.74364662e-01 -1.23385847e+00 4.73023802e-02 9.43654358e-01
3.97301108e-01 -8.14992785e-01 1.26832950e+00 -2.53520966e-01
-4.61505413e-01 6.46956801e-01 5.83678782e-01 5.03620684e-01
3.12236905e-01 -1.79624856e+00 3.26900870e-01 8.03794622e-01
5.59060514e-01 5.74183583e-01 1.08709979e+00 -5.54374218e-01
-1.16454327e+00 -1.35766280e+00 1.51833728e-01 -8.72259855e-01
4.68968362e-01 -6.42713785e-01 -7.56517589e-01 1.30070913e+00
7.89001644e-01 6.95870221e-02 1.06235909e+00 -3.97253484e-01
-3.93491715e-01 1.77141815e-01 -1.53571653e+00 4.61109072e-01
5.53504884e-01 -1.00485456e+00 -6.49983644e-01 3.12395036e-01
7.49333024e-01 -4.99709696e-01 -1.23522985e+00 -2.44647518e-01
4.33844984e-01 -1.02234936e+00 1.24828887e+00 -7.50674427e-01
8.71277928e-01 -1.89981222e-01 -4.92705375e-01 -1.74846351e+00
-3.48895431e-01 -7.03123629e-01 -6.49312973e-01 1.32785583e+00
9.43676829e-02 -4.87371385e-01 7.41363466e-01 4.44537193e-01
-1.18334861e-02 -3.86940688e-01 -8.15480590e-01 -4.70667392e-01
3.55318576e-01 -3.46840233e-01 1.01748204e+00 9.80535626e-01
-5.40304482e-01 1.46999419e-01 -5.52858114e-01 3.17351967e-01
8.94511819e-01 -1.87003896e-01 8.13579440e-01 -1.03771365e+00
-8.08152020e-01 -1.61788866e-01 -6.76875055e-01 -5.66069722e-01
4.60642189e-01 -9.77857649e-01 -1.57850951e-01 -1.15451705e+00
-2.01093778e-02 1.58314466e-01 8.32694843e-02 1.88378960e-01
1.76465243e-01 3.26926589e-01 -9.00170952e-02 1.49493784e-01
-5.62944030e-03 7.95762539e-01 1.01027203e+00 -2.76784003e-01
1.06052868e-01 -2.42818043e-01 -5.64482868e-01 8.83070767e-01
6.27478302e-01 -6.93206549e-01 -4.73713726e-01 -1.85528368e-01
-3.50070260e-02 -1.87139496e-01 6.67126179e-01 -7.87668169e-01
2.05738921e-04 1.42614320e-01 1.03342617e+00 -5.08489132e-01
5.76847494e-01 -9.98132169e-01 6.10305011e-01 1.22860909e-01
-2.88328290e-01 -1.70481026e-01 1.03228912e-01 7.32916117e-01
-1.65963784e-01 -1.36397958e-01 9.46337819e-01 -8.53806138e-02
-2.71432161e-01 2.21077010e-01 -5.32027602e-01 1.49125427e-01
9.88844216e-01 -6.19958192e-02 -1.46170080e-01 -6.95542157e-01
-1.22279823e+00 -2.15299830e-01 5.52979410e-01 1.54433414e-01
7.72268057e-01 -1.64224339e+00 -7.61785269e-01 7.90038824e-01
-1.88682631e-01 1.05329324e-02 2.18103305e-01 1.83573678e-01
-5.22856772e-01 5.79192005e-02 -4.44539189e-01 -6.50756061e-01
-8.63176584e-01 7.68945634e-01 5.87879062e-01 -8.10273811e-02
-4.17851180e-01 9.27645564e-01 3.88411194e-01 -4.16238457e-01
5.96663728e-02 -3.73618081e-02 -3.09263952e-02 -1.97704092e-01
3.78405690e-01 2.38192976e-01 -3.25671196e-01 -7.51942039e-01
-7.60928169e-02 7.64573887e-02 8.55190009e-02 -4.26354617e-01
1.48957229e+00 1.27671838e-01 -4.49046418e-02 3.29749227e-01
1.61339915e+00 -1.17915347e-01 -1.61172199e+00 -2.90839896e-02
-7.53456056e-01 -1.08111180e-01 1.87779799e-01 -4.41417217e-01
-1.26067519e+00 9.36870396e-01 6.57336056e-01 4.23700631e-01
1.00030935e+00 3.33297044e-01 2.94917285e-01 1.76392749e-01
-1.38937697e-01 -8.60210121e-01 1.21531941e-01 2.91134417e-01
1.17905450e+00 -9.58171487e-01 -3.32784384e-01 5.87803200e-02
-7.99209654e-01 9.42348599e-01 2.41097748e-01 -8.68668258e-01
1.03829873e+00 5.94357610e-01 -2.05720365e-01 -4.09231275e-01
-5.45697808e-01 3.13131772e-02 5.00738502e-01 1.16401589e+00
-3.07322834e-02 2.83966333e-01 7.24411547e-01 6.09925628e-01
-8.21076632e-01 -3.15865278e-01 3.72054517e-01 5.83125651e-01
5.04214987e-02 -8.65453959e-01 -6.13065541e-01 3.33557427e-01
-4.72281486e-01 -9.23759565e-02 -6.85898125e-01 6.92629993e-01
2.82786727e-01 5.51418841e-01 5.36923528e-01 -5.36821723e-01
-1.74230218e-01 3.85454714e-01 5.50064921e-01 -5.54038227e-01
-2.56672502e-01 2.54233420e-01 -3.58523548e-01 -3.63924086e-01
-3.46550703e-01 -7.11289227e-01 -1.15385389e+00 -4.04489607e-01
5.54050356e-02 1.22152850e-01 6.42735481e-01 9.19321716e-01
2.94951707e-01 7.34939575e-01 4.67343062e-01 -6.68193877e-01
-5.42749703e-01 -1.19895625e+00 -8.33357811e-01 5.39547741e-01
3.51078182e-01 -4.07211095e-01 -4.66737837e-01 7.65997231e-01]
|
[11.566797256469727, -0.07775796949863434]
|
3a3123d2-8455-4d86-9f10-f95d2936c342
|
case-aligning-coarse-to-fine-cognition-and
|
2208.08845
| null |
https://arxiv.org/abs/2208.08845v2
|
https://arxiv.org/pdf/2208.08845v2.pdf
|
CASE: Aligning Coarse-to-Fine Cognition and Affection for Empathetic Response Generation
|
Empathetic conversation is psychologically supposed to be the result of conscious alignment and interaction between the cognition and affection of empathy. However, existing empathetic dialogue models usually consider only the affective aspect or treat cognition and affection in isolation, which limits the capability of empathetic response generation. In this work, we propose the CASE model for empathetic dialogue generation. It first builds upon a commonsense cognition graph and an emotional concept graph and then aligns the user's cognition and affection at both the coarse-grained and fine-grained levels. Through automatic and manual evaluation, we demonstrate that CASE outperforms state-of-the-art baselines of empathetic dialogues and can generate more empathetic and informative responses.
|
['Minlie Huang', 'Zheng Zhang', 'Bo wang', 'Chujie Zheng', 'Jinfeng Zhou']
|
2022-08-18
| null | null | null | null |
['empathetic-response-generation']
|
['natural-language-processing']
|
[-3.39794487e-01 7.30246544e-01 2.99537033e-02 -5.17615080e-01
-1.55643240e-01 -5.41417956e-01 8.31782639e-01 1.54409915e-01
-1.15794621e-01 7.98034549e-01 9.86656010e-01 2.34812260e-01
1.86980233e-01 -8.33511472e-01 3.67443323e-01 -2.60079205e-01
4.29943532e-01 6.25637174e-01 -5.58880270e-01 -9.31463659e-01
1.48639277e-01 1.24774508e-01 -1.00330877e+00 7.40322232e-01
9.94623065e-01 6.03059769e-01 -4.78307009e-01 7.63225496e-01
-2.19472229e-01 1.61685944e+00 -9.04668510e-01 -9.66150463e-01
-2.44595006e-01 -1.17786443e+00 -1.44338107e+00 -3.68672311e-01
-2.53346890e-01 -3.99289727e-01 -1.59484837e-02 8.46029222e-01
6.59206629e-01 4.11778867e-01 8.05936575e-01 -1.35779428e+00
-8.22405338e-01 8.70068848e-01 -1.73955411e-01 -1.92857489e-01
8.97764266e-01 2.21560642e-01 1.16965032e+00 -4.72046614e-01
5.72484791e-01 1.51166892e+00 6.88408792e-01 1.22395039e+00
-1.07043076e+00 -5.39370358e-01 -8.16470161e-02 3.44866991e-01
-3.57862025e-01 -1.94317177e-01 1.01028633e+00 -4.45321918e-01
8.96800637e-01 2.31188983e-01 1.12554967e+00 1.54216588e+00
4.70160544e-02 8.56813908e-01 1.21365809e+00 -2.97505230e-01
2.65564293e-01 1.66356742e-01 2.73387820e-01 8.33819509e-02
-6.26751482e-01 8.58718306e-02 -7.49007940e-01 -4.01347876e-01
5.76257765e-01 -5.25049806e-01 -2.41765559e-01 1.69129401e-01
-1.00870001e+00 1.43529534e+00 5.17752230e-01 2.93746829e-01
-9.72502112e-01 -9.66942012e-02 7.10548282e-01 3.64096195e-01
3.83861065e-01 1.17292106e+00 -2.48592235e-02 -8.66005719e-01
-4.72424179e-01 4.29929495e-01 1.45681357e+00 6.39668405e-01
1.74119890e-01 -1.91906303e-01 -6.30243480e-01 1.17699921e+00
3.48321199e-02 3.20450276e-01 5.59362948e-01 -1.27550340e+00
-1.80312395e-01 6.77590251e-01 3.06254894e-01 -8.78761470e-01
-6.48858368e-01 1.88456625e-01 -4.43029106e-01 -3.00152902e-03
2.41297632e-01 -7.45917022e-01 7.30153844e-02 2.10411572e+00
4.66385096e-01 -5.63926637e-01 6.09427035e-01 1.13831794e+00
1.29634321e+00 6.61436379e-01 6.43375158e-01 -3.26929480e-01
1.57936716e+00 -1.25142658e+00 -9.82704401e-01 -3.47867221e-01
6.06583178e-01 -8.09365273e-01 1.17619729e+00 2.40949854e-01
-1.49734902e+00 -2.59751946e-01 -6.33886039e-01 -2.71346956e-01
2.73958556e-02 2.91239657e-02 8.96887779e-01 3.31927598e-01
-7.89586961e-01 3.59957546e-01 1.15065940e-01 -4.92836356e-01
-2.90202387e-02 -1.93427857e-02 -2.18598023e-01 4.20294791e-01
-1.84650230e+00 1.54439163e+00 1.03368446e-01 -1.20805398e-01
-3.12535539e-02 -5.72463691e-01 -7.65163124e-01 1.44967169e-01
-3.04817468e-01 -1.06950808e+00 1.87421632e+00 -1.14521480e+00
-2.31393790e+00 9.92882252e-01 2.99476951e-01 -4.27296758e-01
3.69554043e-01 -3.97111356e-01 -1.78275466e-01 3.97764534e-01
-1.66397482e-01 1.12901664e+00 3.36925298e-01 -1.12664390e+00
-2.31858477e-01 1.10436790e-01 4.28769529e-01 7.75214851e-01
-2.77186245e-01 2.84413904e-01 3.57029408e-01 -5.07822692e-01
-4.46225971e-01 -7.47574151e-01 -2.66188942e-02 -3.16254675e-01
-1.98669687e-01 -4.44187284e-01 4.58694667e-01 -4.87022758e-01
9.98727083e-01 -1.93355978e+00 1.58845723e-01 -1.27493143e-01
4.05621052e-01 1.15011640e-01 -8.34361985e-02 1.07844353e+00
-3.18677783e-01 -3.62749189e-01 3.17277133e-01 -3.69241476e-01
4.76968765e-01 3.49078327e-02 -5.28355718e-01 5.91012128e-02
-4.99290004e-02 1.10015476e+00 -1.22740841e+00 -6.59690678e-01
1.91731781e-01 3.39888722e-01 -7.77548969e-01 8.51275861e-01
-2.36340031e-01 5.01710117e-01 -4.11111921e-01 1.99290402e-02
2.55105793e-01 5.91216376e-03 1.35106146e-01 -8.09055716e-02
8.64224881e-02 6.75235510e-01 -1.37216792e-01 1.34294426e+00
-7.21617937e-01 4.94233459e-01 -3.47479875e-03 -2.34076992e-01
1.23253894e+00 6.65001869e-01 3.52671713e-01 -7.46748984e-01
4.69723910e-01 -2.79745281e-01 3.19873512e-01 -4.92765039e-01
7.54038990e-01 -1.02378595e+00 -6.20869040e-01 1.22266281e+00
-1.17292330e-01 -6.11424208e-01 -2.36586496e-01 6.07034385e-01
8.24251473e-01 1.40069947e-01 6.77595794e-01 -6.35533556e-02
4.78081524e-01 1.22627564e-01 1.79665387e-01 3.67153019e-01
-5.37195802e-01 1.12504028e-01 9.57282782e-01 -2.59164035e-01
-5.67328691e-01 -7.95703053e-01 2.53485918e-01 1.53336358e+00
2.68520087e-01 -3.18052799e-01 -1.06559122e+00 -3.39042395e-01
-3.94191414e-01 1.44702804e+00 -8.94554794e-01 -6.37063026e-01
-3.54651898e-01 -4.76500392e-01 8.42541277e-01 3.07934940e-01
5.91530025e-01 -1.82494211e+00 -8.44262421e-01 3.77527386e-01
-8.68273616e-01 -9.71517563e-01 -1.96477845e-01 -1.68692768e-01
-3.50300878e-01 -7.46979654e-01 -1.31060511e-01 -4.30594265e-01
1.74463496e-01 3.12297046e-02 1.36100161e+00 2.43037380e-02
1.02143884e-01 4.89971787e-01 -7.77994633e-01 -5.08489132e-01
-9.05931115e-01 -3.12388510e-01 -1.85095802e-01 -2.26272374e-01
5.17423153e-01 -7.15037167e-01 -6.88194633e-01 9.00850147e-02
-4.13344324e-01 5.82532585e-01 1.58898607e-01 8.71146977e-01
-3.85570705e-01 -7.35815644e-01 1.18694806e+00 -7.92139530e-01
1.70679760e+00 -7.43735909e-01 7.06496775e-01 1.11621246e-02
-2.41250932e-01 -2.74380058e-01 7.32274413e-01 -5.66779196e-01
-1.72023821e+00 -3.79397631e-01 -3.53094369e-01 1.77556966e-02
-2.53502160e-01 2.58172572e-01 2.14974899e-02 2.00661525e-01
1.00910044e+00 -1.82697445e-01 1.94434375e-01 9.48556364e-02
9.70973372e-01 9.15965855e-01 8.08764756e-01 -1.21690869e+00
2.44501173e-01 1.64696455e-01 -5.18698156e-01 -5.23718178e-01
-1.13526583e+00 -3.95498902e-01 -2.97844797e-01 -8.25467587e-01
9.85223234e-01 -8.39057863e-01 -1.23752546e+00 2.16339111e-01
-1.65756202e+00 -4.69955653e-01 -5.72271228e-01 3.59214395e-01
-1.06920707e+00 3.88001561e-01 -1.26324558e+00 -7.79536486e-01
-1.00017357e+00 -5.13810396e-01 7.47964978e-01 3.63760263e-01
-1.61218238e+00 -1.08272767e+00 5.07903337e-01 6.18030131e-01
4.38395649e-01 3.36517602e-01 1.09876287e+00 -1.01166606e+00
7.18957603e-01 -3.42729956e-01 -1.42389104e-01 9.93426442e-02
-8.35996568e-02 -2.64694154e-01 -1.05444229e+00 4.12724227e-01
5.05305886e-01 -1.11958265e+00 9.00772735e-02 -1.79970190e-01
3.12079042e-01 -6.59187198e-01 2.46238142e-01 -4.21080440e-02
5.09633839e-01 3.38642262e-02 9.65156078e-01 6.92879558e-02
1.96851030e-01 1.28292739e+00 9.91838574e-01 9.93786216e-01
7.48978198e-01 5.45391917e-01 2.03508139e-01 -1.86377943e-01
2.40145355e-01 -2.58707345e-01 4.80165511e-01 7.07037687e-01
-3.81641723e-02 -3.85875292e-02 -5.97773969e-01 3.21154803e-01
-2.16259098e+00 -1.47190833e+00 -3.46013784e-01 1.41480398e+00
1.52290881e+00 -5.10961831e-01 2.61499494e-01 -3.52868140e-01
6.03676617e-01 2.02479258e-01 -1.76898271e-01 -1.43100858e+00
1.21043192e-03 1.88403189e-01 -6.12717509e-01 7.78408408e-01
-6.03530824e-01 1.37428164e+00 6.15477467e+00 4.09320801e-01
-7.85448432e-01 2.17943285e-02 4.09337819e-01 -3.16287994e-01
-2.92691737e-01 -2.33202994e-01 -5.45736216e-02 1.08784132e-01
6.25250399e-01 -4.34004396e-01 2.48816222e-01 9.51372623e-01
2.06188157e-01 2.09664311e-02 -1.42135024e+00 8.95645857e-01
6.00777194e-02 -7.33675659e-01 -8.39852616e-02 -2.63041317e-01
4.23539221e-01 -7.46818602e-01 -2.65414655e-01 7.22600043e-01
7.47083724e-01 -1.21603847e+00 6.62725329e-01 6.14275813e-01
2.63862401e-01 -7.32752979e-01 9.03758645e-01 3.82431746e-01
-4.39550757e-01 1.14520051e-01 -1.10539701e-02 -5.44462681e-01
4.76752430e-01 -1.80380002e-01 -8.89133215e-01 3.63631523e-04
1.49652779e-01 4.43229228e-01 2.93889139e-02 3.83187115e-01
-8.03829730e-01 3.11747581e-01 1.94505453e-01 -3.71260405e-01
3.30817968e-01 -4.25202191e-01 4.56266016e-01 1.45098317e+00
-1.14368692e-01 7.72034526e-01 -6.06314689e-02 1.11172664e+00
4.79372069e-02 4.66603726e-01 -2.74738580e-01 5.57374617e-04
6.65693641e-01 1.69134629e+00 6.60983771e-02 -3.78371924e-01
2.28231847e-02 1.12876594e+00 5.26148438e-01 2.25355905e-02
-1.03588092e+00 3.47728282e-02 7.54316926e-01 -5.75062096e-01
-3.90559614e-01 4.22654212e-01 -5.19533634e-01 -7.60587156e-01
-5.96909344e-01 -1.07982349e+00 4.52129483e-01 -1.24387884e+00
-1.61242414e+00 5.70944130e-01 -1.67857185e-01 -8.15175056e-01
-7.83198953e-01 -2.89574087e-01 -1.28300834e+00 1.03191996e+00
-7.97712088e-01 -1.34442949e+00 -4.49938983e-01 5.23219287e-01
3.12947124e-01 3.25360805e-01 1.48078156e+00 -3.85680526e-01
-1.87398702e-01 4.39940691e-01 -8.90523612e-01 4.28070165e-02
1.13638723e+00 -1.26397860e+00 3.36018689e-02 -1.30254868e-02
-6.48696721e-01 6.28559470e-01 1.17719138e+00 -3.43284100e-01
-7.51149833e-01 -5.71850479e-01 1.13361037e+00 -4.72490549e-01
9.73001659e-01 1.00702353e-01 -7.95020103e-01 4.00815934e-01
9.40028548e-01 -9.72211719e-01 1.39415097e+00 3.74230623e-01
-4.57817584e-01 6.44156873e-01 -1.41350996e+00 1.05166161e+00
7.01556623e-01 -5.62002301e-01 -1.34091127e+00 4.24955934e-01
5.42487919e-01 -2.75569648e-01 -1.18868399e+00 -1.19038410e-01
5.33396780e-01 -1.28584874e+00 8.66250277e-01 -9.48681831e-01
1.10933506e+00 4.20770228e-01 1.22452103e-01 -1.66406310e+00
-2.51929373e-01 -9.95287776e-01 -2.73524988e-02 1.19314253e+00
8.57606605e-02 -3.81008029e-01 4.34016377e-01 1.09062290e+00
-2.03274563e-01 -5.67140341e-01 -6.04484320e-01 -8.19286630e-02
4.27903861e-01 -1.67908862e-01 5.69180191e-01 1.39285529e+00
1.55671620e+00 1.38203895e+00 -6.00959361e-01 -4.98144150e-01
1.55332431e-01 4.64095265e-01 9.34018314e-01 -1.36505675e+00
-4.32437450e-01 -1.01989853e+00 1.52562559e-01 -8.28738093e-01
5.97478449e-01 -5.73977768e-01 2.41991103e-01 -1.60172391e+00
3.63253623e-01 4.83218255e-03 5.32195926e-01 4.91322249e-01
-3.40690851e-01 -1.92313030e-01 2.65496105e-01 1.38482273e-01
-5.77055871e-01 9.85606432e-01 1.37704301e+00 4.50587481e-01
-4.17131960e-01 -3.25111836e-01 -1.20194876e+00 8.97491574e-01
1.17829132e+00 -1.27806827e-01 -3.57647955e-01 2.72133559e-01
3.64136249e-01 6.27549767e-01 3.61431569e-01 -4.00889307e-01
2.77658701e-01 -4.84489024e-01 7.05490112e-02 -1.66952208e-01
9.06894326e-01 -2.43873462e-01 -2.90506035e-01 3.17138314e-01
-8.68942082e-01 -1.11246593e-01 6.02335595e-02 9.45980251e-02
-3.22588265e-01 -2.64937103e-01 1.05882692e+00 -1.74812347e-01
-2.95503259e-01 -2.58950084e-01 -5.97455621e-01 1.93685845e-01
9.46005821e-01 -1.67781606e-01 -6.91431522e-01 -1.26986766e+00
-6.93915367e-01 3.23433876e-01 4.49638277e-01 4.04648572e-01
5.34795165e-01 -1.35910010e+00 -7.99502969e-01 -5.58556497e-01
2.32032061e-01 -4.96651769e-01 5.66069663e-01 8.73808384e-01
-2.45062634e-01 5.58331050e-02 -6.33211017e-01 1.71048254e-01
-1.36982298e+00 3.10699731e-01 5.59950948e-01 -3.07973325e-01
-5.68963110e-01 8.71548831e-01 4.09235954e-01 -7.02829540e-01
2.55440772e-02 3.79325390e-01 -7.43586779e-01 3.16154271e-01
6.02406442e-01 2.67660826e-01 -7.21882164e-01 -8.05287659e-01
2.38688905e-02 -1.68810785e-02 6.86240345e-02 -4.69339252e-01
8.91871691e-01 -4.14713696e-02 -5.28682351e-01 3.23618948e-01
6.76943064e-01 -9.57399979e-02 -8.18645298e-01 8.60853791e-02
-2.98862547e-01 -5.80671802e-02 -3.96887481e-01 -1.19977045e+00
-2.83282220e-01 9.35306549e-01 -3.91358078e-01 3.08077753e-01
8.71000171e-01 5.13939783e-02 1.08248091e+00 4.21890318e-01
7.20381662e-02 -1.44445550e+00 6.16436660e-01 8.19953740e-01
1.30063951e+00 -7.90689945e-01 -1.64424405e-01 -4.13954407e-01
-1.63396049e+00 1.23373854e+00 1.11740136e+00 4.70307693e-02
1.11692538e-02 7.60084167e-02 6.36486769e-01 -4.00812596e-01
-1.28858876e+00 1.29058212e-01 5.81198707e-02 5.72533429e-01
9.29245889e-01 3.59841406e-01 -6.71092689e-01 1.32678914e+00
-9.99311626e-01 -2.39968568e-01 6.78123713e-01 5.47184408e-01
-3.84789407e-01 -9.67479527e-01 -1.82788521e-01 -1.17349893e-01
-1.33168235e-01 -1.57192022e-01 -1.52187705e+00 6.92649901e-01
-2.89532155e-01 1.31852138e+00 -8.02693442e-02 -4.42833722e-01
4.18394417e-01 2.36015707e-01 4.59996581e-01 -6.12116575e-01
-1.50901043e+00 -3.93655747e-01 8.58135641e-01 -5.43603480e-01
-4.11439925e-01 -5.93631089e-01 -1.68588793e+00 -8.17273200e-01
9.33432300e-03 4.68883544e-01 2.46016428e-01 1.08547068e+00
1.43092781e-01 1.55677870e-01 6.21532023e-01 -6.63912177e-01
-8.82858574e-01 -1.40255857e+00 -3.38047296e-01 9.33276296e-01
-3.72171253e-01 -2.63771325e-01 -3.75740856e-01 -1.99958116e-01]
|
[13.137838363647461, 7.656383514404297]
|
45d7ed43-444e-4a61-9b36-46d6cb9f9fe4
|
show-me-what-i-like-detecting-user-specific
|
2207.08352
| null |
https://arxiv.org/abs/2207.08352v2
|
https://arxiv.org/pdf/2207.08352v2.pdf
|
Show Me What I Like: Detecting User-Specific Video Highlights Using Content-Based Multi-Head Attention
|
We propose a method to detect individualized highlights for users on given target videos based on their preferred highlight clips marked on previous videos they have watched. Our method explicitly leverages the contents of both the preferred clips and the target videos using pre-trained features for the objects and the human activities. We design a multi-head attention mechanism to adaptively weigh the preferred clips based on their object- and human-activity-based contents, and fuse them using these weights into a single feature representation for each user. We compute similarities between these per-user feature representations and the per-frame features computed from the desired target videos to estimate the user-specific highlight clips from the target videos. We test our method on a large-scale highlight detection dataset containing the annotated highlights of individual users. Compared to current baselines, we observe an absolute improvement of 2-4% in the mean average precision of the detected highlights. We also perform extensive ablation experiments on the number of preferred highlight clips associated with each user as well as on the object- and human-activity-based feature representations to validate that our method is indeed both content-based and user-specific.
|
['Dinesh Manocha', 'Viswanathan Swaminathan', 'Stefano Petrangeli', 'Gang Wu', 'Uttaran Bhattacharya']
|
2022-07-18
| null | null | null | null |
['highlight-detection']
|
['computer-vision']
|
[ 1.97560400e-01 -3.70562494e-01 -2.77477890e-01 -2.85722464e-01
-1.21915758e+00 -8.01072538e-01 5.46916425e-01 3.04437906e-01
-4.77828920e-01 1.06418744e-01 7.57689238e-01 6.31263316e-01
1.96074381e-01 -1.39664516e-01 -5.49017727e-01 -3.45474571e-01
-5.49799323e-01 -3.37434709e-01 4.60279256e-01 2.26598263e-01
5.25372386e-01 2.75327653e-01 -1.96802986e+00 8.75186145e-01
5.44346511e-01 1.38286424e+00 8.28372240e-02 1.00246489e+00
4.37423110e-01 7.25492954e-01 -7.30840445e-01 -1.20802477e-01
5.58157146e-01 -3.18016618e-01 -1.74235880e-01 3.26178581e-01
1.37382054e+00 -5.41796744e-01 -6.07103229e-01 7.75708854e-01
5.54146051e-01 3.50390196e-01 3.90805960e-01 -1.24136746e+00
-3.86224985e-01 3.36918950e-01 -9.45864975e-01 8.63082528e-01
7.87666142e-01 2.81890541e-01 1.21274269e+00 -1.20616436e+00
8.62697482e-01 9.85183597e-01 5.41464269e-01 8.86730924e-02
-9.23347175e-01 -7.78052032e-01 4.52744752e-01 2.72968948e-01
-1.36861718e+00 -7.78715312e-01 6.73432410e-01 -6.54379189e-01
6.75297737e-01 3.57317895e-01 8.62277925e-01 8.57920706e-01
-1.27097974e-02 9.34338450e-01 5.18174469e-01 -2.25089461e-01
-4.85181883e-02 1.32888630e-01 -9.43061188e-02 7.87050843e-01
1.95135564e-01 -1.34196758e-01 -1.21167552e+00 -2.59641916e-01
6.50782049e-01 2.12590307e-01 -5.73679745e-01 -3.25153440e-01
-1.41496599e+00 6.71044171e-01 1.87628999e-01 1.65729567e-01
-5.36233425e-01 2.30991676e-01 5.22476256e-01 -1.87874645e-01
5.13752401e-01 6.30263925e-01 -4.51335728e-01 -4.08332199e-01
-1.39255226e+00 3.35505486e-01 4.05479997e-01 1.19061863e+00
5.48316061e-01 -1.13688104e-01 -9.64757681e-01 4.05772746e-01
-1.73571065e-01 5.02764821e-01 2.26219341e-01 -1.14977407e+00
3.62855524e-01 6.86610758e-01 5.23401022e-01 -1.36637402e+00
-1.27480969e-01 -2.37066090e-01 1.89743206e-01 -9.81104076e-02
3.62681597e-01 -3.49168986e-01 -7.85956204e-01 1.84668124e+00
2.44651705e-01 5.15516162e-01 -7.01642573e-01 1.10093665e+00
6.62777007e-01 5.24739265e-01 2.54100412e-01 -2.08287820e-01
1.67347443e+00 -1.08352590e+00 -5.08172810e-01 -1.32223070e-01
2.75665492e-01 -8.71753633e-01 1.26871407e+00 3.97477537e-01
-1.26444101e+00 -7.97735989e-01 -8.47517848e-01 -4.72187363e-02
-3.65023762e-02 5.51562190e-01 3.27198416e-01 4.21502709e-01
-8.07755649e-01 4.91132438e-01 -3.92041981e-01 -3.50254387e-01
3.65777910e-01 9.62600559e-02 -2.54913688e-01 2.99771249e-01
-7.37214267e-01 5.19823372e-01 1.33185238e-01 -2.99277693e-01
-1.11345935e+00 -1.21891308e+00 -6.89793468e-01 2.29518309e-01
6.48290575e-01 -1.87306225e-01 1.33226275e+00 -1.24264407e+00
-8.78294706e-01 9.00687397e-01 -3.57747287e-01 -1.67076141e-01
3.97664845e-01 -8.92655134e-01 -3.92536044e-01 7.51929343e-01
1.89819574e-01 7.49385178e-01 1.13933182e+00 -1.09603345e+00
-1.37301052e+00 1.13762163e-01 3.21585894e-01 3.57818782e-01
-7.92024672e-01 3.47103000e-01 -1.04364657e+00 -7.80433953e-01
-4.35504854e-01 -7.97604620e-01 3.37858260e-01 4.33397084e-01
-4.18318242e-01 1.83617212e-02 9.47312653e-01 -8.69794846e-01
1.56743300e+00 -2.21233940e+00 -4.62051742e-02 2.74229795e-01
5.18878162e-01 -9.63881686e-02 -3.47550750e-01 2.81104535e-01
7.69794956e-02 -1.87191620e-01 4.73901153e-01 -2.21577242e-01
-2.50192523e-01 -4.97569889e-01 -4.23545599e-01 5.31287372e-01
1.75065055e-01 5.57009757e-01 -1.14683330e+00 -5.73016286e-01
8.47851634e-02 6.33268893e-01 -7.25790024e-01 4.92583394e-01
-7.41969272e-02 1.35295674e-01 -2.58749098e-01 7.54957199e-01
2.88789153e-01 -8.54197368e-02 1.21964082e-01 -7.11696684e-01
-2.24151090e-01 1.40755959e-02 -1.10054088e+00 1.59732842e+00
-2.47444153e-01 9.92377877e-01 -5.71400449e-02 -2.49020979e-01
5.17977178e-01 2.82379419e-01 6.83046043e-01 -5.22343814e-01
-1.75747443e-02 -1.89291865e-01 -1.29645318e-01 -6.90814614e-01
8.86674643e-01 4.56067830e-01 -1.64378777e-01 5.23371756e-01
1.55444458e-01 5.24711132e-01 4.57828879e-01 5.35803735e-01
1.25014055e+00 2.64964044e-01 3.14945787e-01 -2.12934688e-01
1.86862275e-01 -3.73838276e-01 3.98789972e-01 8.55571926e-01
-4.57441509e-01 8.77624989e-01 5.90544045e-01 -4.47704971e-01
-9.54062104e-01 -9.00868058e-01 3.96982253e-01 2.03504419e+00
1.12182766e-01 -9.16633368e-01 -7.70538986e-01 -8.06458116e-01
1.39297754e-01 4.63261813e-01 -9.91204917e-01 -1.75629720e-01
-5.12286782e-01 1.70391575e-02 3.29125941e-01 5.47833681e-01
1.59828752e-01 -1.03425169e+00 -1.18866003e+00 -6.71786582e-03
-2.55677313e-01 -1.13200915e+00 -1.42239797e+00 -2.74176300e-01
-1.89939231e-01 -1.31691730e+00 -6.84402347e-01 -4.51366872e-01
6.77621841e-01 7.35942364e-01 1.11657107e+00 1.71817482e-01
-4.29658175e-01 9.45913553e-01 -3.88008207e-01 -2.08825439e-01
2.64067590e-01 -2.83601731e-01 1.17578976e-01 5.06239653e-01
2.29018554e-01 -2.39169866e-01 -1.13070095e+00 2.21827537e-01
-6.80790663e-01 -5.58270141e-02 3.55978310e-01 3.15214097e-01
3.42859507e-01 -1.94684774e-01 -8.71552038e-04 -6.56330347e-01
5.45346797e-01 -3.33331257e-01 -2.54795879e-01 2.51262844e-01
-5.87937981e-02 -3.72941703e-01 2.23186985e-01 -8.65076125e-01
-8.11424434e-01 1.87853396e-01 6.61612630e-01 -8.85234833e-01
1.80945337e-01 2.12935776e-01 4.26940583e-02 -1.85787808e-02
6.91539347e-01 -1.21724762e-01 -5.72768450e-01 -2.32464582e-01
3.68891716e-01 3.73829424e-01 7.92617440e-01 -5.07147908e-01
7.77790010e-01 4.58542049e-01 -4.57914084e-01 -5.67712665e-01
-1.30270827e+00 -8.51919055e-01 -5.19072235e-01 -5.46040654e-01
7.95836866e-01 -1.31017447e+00 -7.23933816e-01 1.53960586e-01
-1.09926176e+00 -1.96194485e-01 -1.82719991e-01 3.32211316e-01
-5.01705468e-01 3.82946074e-01 -4.86033887e-01 -9.23994601e-01
-4.79116201e-01 -1.00012195e+00 1.46100092e+00 3.89927238e-01
-6.68665886e-01 -5.83830714e-01 -7.66984597e-02 -6.95475116e-02
3.46093476e-01 4.86481637e-01 3.17496508e-01 -5.86605132e-01
-5.40760100e-01 -7.08387554e-01 -5.13219237e-01 -1.51592895e-01
8.56080726e-02 4.77089763e-01 -1.08156896e+00 -4.83603507e-01
-3.67815733e-01 -3.20313498e-02 9.01793480e-01 5.16539276e-01
1.10547042e+00 -3.45960200e-01 -4.01431918e-01 3.07533056e-01
1.34197366e+00 -1.14946306e-01 2.47402862e-01 5.44470213e-02
7.21475005e-01 3.48468810e-01 1.01485097e+00 9.57978368e-01
7.67005607e-02 9.37446296e-01 2.94189513e-01 -1.12259708e-01
-1.10835969e-01 -2.64664978e-01 6.94049716e-01 1.88049749e-01
-4.16931868e-01 -7.61747658e-02 -4.37014133e-01 7.06467927e-01
-1.81366014e+00 -1.53666341e+00 2.95457244e-01 2.51597095e+00
7.37005711e-01 3.87713015e-01 7.94874847e-01 -2.51536191e-01
9.79636967e-01 3.50756764e-01 -4.50102299e-01 8.25767964e-02
8.26537088e-02 1.45492196e-01 3.73795688e-01 1.34752229e-01
-1.34387970e+00 6.73691273e-01 6.72271633e+00 5.13014913e-01
-1.18482077e+00 1.67892724e-01 4.78519201e-01 -9.71716404e-01
8.68115872e-02 -2.41125301e-01 -8.31157982e-01 8.49509656e-01
9.76505339e-01 -2.64003813e-01 3.72190803e-01 1.04052246e+00
5.75331569e-01 -3.51849914e-01 -1.27618885e+00 1.02445924e+00
4.41634595e-01 -1.37249148e+00 -8.06214213e-02 6.71379641e-02
7.45393634e-01 -2.43112236e-01 3.48113954e-01 3.46445143e-01
-2.88421273e-01 -5.24383903e-01 1.17180502e+00 5.85382640e-01
9.52619493e-01 -7.05408812e-01 1.71749726e-01 -2.98205376e-01
-1.48403060e+00 -3.57576847e-01 -1.05695374e-01 -4.09571156e-02
6.38954267e-02 3.85705441e-01 -6.00295007e-01 -1.75646171e-01
8.91474068e-01 7.91565716e-01 -7.47681141e-01 1.40065217e+00
-9.16639566e-02 6.82584465e-01 -2.30435789e-01 2.33686373e-01
2.41222093e-03 3.54157448e-01 4.78471577e-01 1.73769212e+00
4.90625679e-01 -7.42831677e-02 4.59761322e-01 6.29414201e-01
-2.12364823e-01 1.88575462e-01 -1.56017348e-01 -1.64484382e-01
6.27570331e-01 1.77965915e+00 -6.85927510e-01 -7.47000277e-01
-5.74033916e-01 1.00955892e+00 2.43508607e-01 3.15161914e-01
-1.17943561e+00 -5.89672208e-01 7.60035515e-01 4.09571588e-01
7.19450653e-01 -1.05575152e-01 2.98638225e-01 -1.23362458e+00
1.59625605e-01 -8.12354088e-01 5.96331835e-01 -9.36912775e-01
-9.84565675e-01 5.66445172e-01 -8.51282850e-03 -1.39478385e+00
-1.14258565e-02 -3.31550956e-01 -1.01612747e+00 6.08607173e-01
-1.10628080e+00 -1.08661902e+00 -8.60955596e-01 5.49639523e-01
7.65306950e-01 2.76489649e-02 5.98278642e-01 4.39536631e-01
-4.86966461e-01 8.04944575e-01 -3.17502856e-01 2.26731688e-01
1.20016885e+00 -1.05477464e+00 2.09166959e-01 9.98916268e-01
3.08686286e-01 6.42753541e-01 9.22801077e-01 -8.00783813e-01
-1.25651228e+00 -8.88240576e-01 4.42762762e-01 -8.62523973e-01
6.19885623e-01 -3.73048127e-01 -7.00107217e-01 7.01706946e-01
2.92007506e-01 3.22650492e-01 8.86978090e-01 2.61581212e-01
-5.73126495e-01 8.72338936e-03 -8.55022609e-01 5.59705317e-01
1.15864408e+00 -6.28792644e-01 -3.09690863e-01 3.25655520e-01
5.63968956e-01 -6.67339981e-01 -7.37329781e-01 1.94866564e-02
9.61299241e-01 -1.12347519e+00 9.13302839e-01 -7.11857200e-01
4.23118502e-01 -3.24326426e-01 -1.28032222e-01 -1.02514660e+00
-5.92728078e-01 -7.39267170e-01 -5.35653293e-01 9.99753058e-01
5.72230667e-02 3.48520070e-01 6.90768361e-01 8.29934180e-01
-1.85340568e-01 -6.30851030e-01 -6.01128101e-01 -2.87366271e-01
-9.86720622e-01 -2.52934903e-01 1.75559595e-01 6.80627167e-01
2.36423612e-01 7.30353668e-02 -8.09856057e-01 1.19105980e-01
4.13088739e-01 2.58902341e-01 9.01951194e-01 -1.08295476e+00
-4.66267377e-01 -3.72154534e-01 -4.07370329e-01 -7.50179291e-01
-1.81099355e-01 -3.49669814e-01 1.13198407e-01 -9.92491066e-01
7.55231798e-01 1.99957177e-01 -7.30864465e-01 5.31316698e-01
-4.66859430e-01 5.75248957e-01 4.01664048e-01 2.12654307e-01
-1.36624742e+00 -3.67824510e-02 8.16189408e-01 2.68324446e-02
-5.73547363e-01 -1.93315297e-01 -8.93209338e-01 7.22650051e-01
3.53632569e-01 -3.29284132e-01 -1.95289582e-01 -3.33344489e-02
7.93189481e-02 -3.09131473e-01 2.93722719e-01 -1.32502699e+00
2.46253703e-02 -2.37216935e-01 8.18589687e-01 -5.73441565e-01
5.71443260e-01 -5.35269201e-01 -3.06165963e-01 1.80632815e-01
-7.03964293e-01 3.77975814e-02 3.25797319e-01 7.21486032e-01
1.63604707e-01 1.36757672e-01 7.00620174e-01 -6.31623343e-03
-9.20568705e-01 3.13847005e-01 -2.40368217e-01 1.82498932e-01
1.12777221e+00 -3.79530460e-01 -2.39237875e-01 -8.15794110e-01
-5.28167963e-01 -3.84434238e-02 4.78043199e-01 5.72441280e-01
6.46947503e-01 -1.23623991e+00 -5.35815537e-01 3.54741653e-03
4.24100041e-01 -7.12240875e-01 2.84951806e-01 9.25047219e-01
-1.01648189e-01 1.60729378e-01 -4.48984653e-01 -3.77018332e-01
-1.55802405e+00 8.46461713e-01 -3.29927430e-02 9.28726122e-02
-5.78252494e-01 9.69453633e-01 3.71445566e-01 6.15221739e-01
5.05577266e-01 -3.27381015e-01 -1.32353723e-01 4.54901010e-01
1.05350828e+00 4.04585630e-01 -3.52911711e-01 -8.49323094e-01
-3.49890769e-01 4.74740118e-01 -2.87321538e-01 -6.70970604e-02
1.13536835e+00 -1.17687009e-01 4.98422921e-01 1.76537082e-01
1.19805336e+00 8.00560474e-01 -1.76830125e+00 -2.29555368e-01
-1.98644921e-01 -8.83452952e-01 -8.85955170e-02 -6.28159642e-01
-1.20898187e+00 5.60764194e-01 6.00888491e-01 6.51657442e-03
1.06302285e+00 -1.31285727e-01 7.00273335e-01 4.92979847e-02
2.29763240e-01 -1.15204179e+00 6.65688813e-01 -8.97116363e-02
9.39315975e-01 -1.08911920e+00 3.04837048e-01 -4.38611031e-01
-8.26050460e-01 1.01671958e+00 6.86410367e-01 -4.33240652e-01
3.48756015e-01 3.05818599e-02 -3.58162560e-02 -2.41577759e-01
-7.84610629e-01 -2.76227236e-01 8.03115785e-01 4.39209640e-01
6.32630825e-01 -1.18849702e-01 1.40859514e-01 5.92683315e-01
1.34393692e-01 -7.05679879e-02 5.48523724e-01 8.37729633e-01
-7.82901525e-01 -4.48420197e-01 -3.42295945e-01 7.61017561e-01
-4.97096777e-01 -3.22360210e-02 -4.15083855e-01 6.07276678e-01
1.07661426e-01 7.05727696e-01 3.61235708e-01 -5.30316532e-01
3.69764507e-01 -1.20637283e-01 4.95189309e-01 -6.85541868e-01
-7.22176552e-01 3.81735891e-01 1.17499016e-01 -9.04105067e-01
-3.80390644e-01 -7.64790535e-01 -1.05338228e+00 8.89826752e-03
-1.38278201e-01 6.93788901e-02 1.80552274e-01 3.46307099e-01
8.02529812e-01 3.70890141e-01 5.14716983e-01 -1.54845107e+00
-1.14908449e-01 -8.19714546e-01 -3.87295187e-01 8.97032678e-01
2.68616080e-01 -1.00218749e+00 -3.71074468e-01 2.67189741e-01]
|
[10.152823448181152, 0.4962579309940338]
|
1f3f6538-0133-4003-b2ba-06c41c96f1e7
|
analyzing-scrna-seq-data-by-ccp-assisted-umap
|
2306.13750
| null |
https://arxiv.org/abs/2306.13750v1
|
https://arxiv.org/pdf/2306.13750v1.pdf
|
Analyzing scRNA-seq data by CCP-assisted UMAP and t-SNE
|
Single-cell RNA sequencing (scRNA-seq) is widely used to reveal heterogeneity in cells, which has given us insights into cell-cell communication, cell differentiation, and differential gene expression. However, analyzing scRNA-seq data is a challenge due to sparsity and the large number of genes involved. Therefore, dimensionality reduction and feature selection are important for removing spurious signals and enhancing downstream analysis. Correlated clustering and projection (CCP) was recently introduced as an effective method for preprocessing scRNA-seq data. CCP utilizes gene-gene correlations to partition the genes and, based on the partition, employs cell-cell interactions to obtain super-genes. Because CCP is a data-domain approach that does not require matrix diagonalization, it can be used in many downstream machine learning tasks. In this work, we utilize CCP as an initialization tool for uniform manifold approximation and projection (UMAP) and t-distributed stochastic neighbor embedding (t-SNE). By using eight publicly available datasets, we have found that CCP significantly improves UMAP and t-SNE visualization and dramatically improve their accuracy.
|
['Gu-Wei Wei', 'Yuta Hozumi']
|
2023-06-23
| null | null | null | null |
['dimensionality-reduction']
|
['methodology']
|
[ 1.74150169e-01 -6.86303079e-01 -7.26079866e-02 1.71894401e-01
-7.82265306e-01 -7.84803152e-01 4.26895231e-01 1.66125476e-01
-1.50117785e-01 6.83658719e-01 5.22603393e-01 -1.50909841e-01
-3.43435258e-01 -5.94911337e-01 -1.77398786e-01 -1.34245837e+00
1.11129113e-01 4.23027366e-01 -1.19212225e-01 1.70394238e-02
2.92068034e-01 6.44829988e-01 -1.24078989e+00 1.36797309e-01
7.19494164e-01 2.61669099e-01 3.36170755e-02 6.98586106e-01
-3.35973561e-01 -9.89394169e-03 -1.68557733e-01 7.89345950e-02
1.69140279e-01 -6.92224801e-01 -3.40743721e-01 -2.27436930e-01
-2.17099458e-01 6.47793949e-01 -1.40721172e-01 8.76095533e-01
8.61614287e-01 2.22310424e-01 8.26089323e-01 -1.38534200e+00
-1.59915119e-01 3.19599807e-01 -7.75842488e-01 -3.96916978e-02
9.46391895e-02 8.74975696e-02 7.31006324e-01 -1.17748523e+00
1.00112760e+00 1.26908731e+00 5.61726809e-01 6.29256546e-01
-1.88084841e+00 -5.04473150e-01 -3.35659325e-01 -9.98797491e-02
-1.69164455e+00 -1.54752076e-01 6.57869875e-01 -7.79467046e-01
4.20167506e-01 5.94911814e-01 8.47588062e-01 8.56078148e-01
1.50287047e-01 4.78625059e-01 1.10286498e+00 -1.97159201e-01
6.18919015e-01 -2.30507106e-01 2.02845395e-01 3.62526447e-01
1.22200117e-01 -2.56684601e-01 -5.15463233e-01 -6.17160380e-01
4.45256293e-01 4.20288682e-01 -5.00327349e-01 -3.84312540e-01
-1.66880703e+00 7.12870121e-01 -3.86954695e-02 4.67161328e-01
-2.10061565e-01 -5.78845367e-02 5.59996068e-01 -3.13834101e-02
2.11097687e-01 5.28429031e-01 -3.73000503e-01 -4.79396671e-01
-7.30454504e-01 -1.67306095e-01 4.89210099e-01 4.71449316e-01
5.67078829e-01 -2.96733290e-01 -4.68022525e-02 1.05105305e+00
-1.00690149e-01 3.44677150e-01 6.36333048e-01 -9.88529563e-01
-1.59429893e-01 8.89183402e-01 -2.00566322e-01 -1.19728112e+00
-6.43334806e-01 -1.26548722e-01 -1.28705585e+00 -1.63114399e-01
5.20936787e-01 -9.58695784e-02 -5.56437373e-01 1.70909429e+00
6.55523241e-01 3.07784885e-01 9.62401778e-02 8.35026145e-01
4.03172582e-01 7.17866063e-01 -3.19650061e-02 -4.45694774e-01
1.19204545e+00 -3.27940106e-01 -7.05695987e-01 5.13530433e-01
9.52534020e-01 -5.06166220e-01 1.07640171e+00 1.59411266e-01
-2.97202080e-01 -6.10592216e-02 -5.70878148e-01 -2.45647669e-01
-4.60524142e-01 1.39224939e-02 3.40284556e-01 3.50407183e-01
-8.11695874e-01 6.16925240e-01 -1.02986026e+00 -6.19386494e-01
6.47494435e-01 2.82261848e-01 -7.17632234e-01 -2.33292639e-01
-6.31571293e-01 4.17288810e-01 1.59354154e-02 -9.01594851e-03
-4.25354987e-01 -9.44313526e-01 -5.65370798e-01 2.42836982e-01
1.08027518e-01 -6.23397171e-01 1.05287626e-01 -3.55783813e-02
-1.62122726e+00 6.73638880e-01 -4.03579026e-01 2.04434454e-01
5.77243567e-02 1.77967936e-01 -1.32389233e-01 1.46446610e-02
-1.03837125e-01 4.84444827e-01 3.27439696e-01 -9.69336152e-01
-2.13838905e-01 -8.30374241e-01 -9.03330922e-01 2.54869834e-02
-4.01050925e-01 -7.76954219e-02 -4.91037339e-01 -4.52942759e-01
6.00606382e-01 -1.19840860e+00 -4.15754676e-01 -4.78161424e-02
-6.36903763e-01 1.07713751e-01 9.60884035e-01 -3.83937865e-01
9.61606503e-01 -2.55124664e+00 7.70890236e-01 3.45115155e-01
3.16777766e-01 9.62463170e-02 -3.34586501e-01 7.46160448e-01
-2.55886793e-01 3.23634595e-01 -3.39885056e-01 -2.19356120e-01
-2.01565966e-01 1.75287593e-02 -1.68430343e-01 7.20339894e-01
-9.74141881e-02 8.21280420e-01 -8.78035843e-01 -4.00892556e-01
2.49131739e-01 8.84135842e-01 -4.30972487e-01 -1.38852030e-01
9.95710939e-02 8.07545364e-01 -2.19314724e-01 6.28204465e-01
6.34909451e-01 -2.82459408e-01 5.89751780e-01 -2.80003905e-01
-2.00824916e-01 -3.42057019e-01 -1.04400325e+00 1.37706327e+00
-1.53728858e-01 7.69542813e-01 -8.32740664e-02 -7.65228629e-01
9.47207749e-01 6.70924932e-02 7.75876224e-01 -7.79642910e-02
2.91021913e-01 7.02029988e-02 -1.65756810e-02 -3.01960766e-01
-4.62275818e-02 -6.10363111e-02 1.29030839e-01 3.92416298e-01
-3.11999828e-01 -2.02576537e-02 2.31271192e-01 3.89847487e-01
1.29720378e+00 -2.24837780e-01 3.09549391e-01 -6.07460737e-01
5.17690361e-01 -5.89787140e-02 1.04752326e+00 1.64608900e-02
-1.02544613e-01 9.71918881e-01 1.02407742e+00 -1.00826189e-01
-8.76237988e-01 -9.02080834e-01 -1.02423087e-01 7.88872361e-01
7.81005919e-02 -4.81006175e-01 -4.44309890e-01 -3.64868939e-01
7.46799037e-02 5.25439143e-01 -6.50848091e-01 -1.46827176e-01
-1.51639819e-01 -1.38216686e+00 4.25399631e-01 3.23345661e-01
-1.71558455e-01 -3.52440596e-01 2.83932984e-02 2.49251053e-02
-1.37778744e-01 -7.04987288e-01 -7.16042459e-01 9.26860645e-02
-9.68773067e-01 -1.36593115e+00 -7.01208472e-01 -4.96638805e-01
1.07124913e+00 4.68909353e-01 2.98149526e-01 -2.46384576e-01
-5.90901375e-01 2.18840167e-01 -2.04479963e-01 -4.40700613e-02
-2.46169835e-01 -1.92861006e-01 3.79496545e-01 7.63772130e-02
4.22161341e-01 -8.34770560e-01 -5.43392897e-01 3.16517323e-01
-9.18819129e-01 3.71975526e-02 4.03487474e-01 1.18554497e+00
9.94408607e-01 1.00844521e-02 4.95383948e-01 -9.55727339e-01
6.19995952e-01 -4.26435649e-01 -7.35611916e-01 8.14660192e-02
-4.93160218e-01 5.55496030e-02 8.39386880e-01 -2.84317970e-01
-6.67045236e-01 2.88609385e-01 1.68184489e-02 -4.13619190e-01
2.87099704e-02 5.37925541e-01 -4.30362463e-01 7.22702444e-02
4.06995326e-01 5.27056992e-01 5.61449468e-01 -4.75355089e-01
2.20336139e-01 6.16105616e-01 6.92444518e-02 -7.85429254e-02
4.88938928e-01 7.63571024e-01 7.45076954e-01 -1.05733371e+00
-2.50489861e-01 -7.39222825e-01 -9.07732010e-01 1.34584889e-01
7.84437418e-01 -7.63803601e-01 -1.06674123e+00 2.30616912e-01
-5.29247701e-01 -1.05928399e-01 -1.06125576e-02 6.26882911e-01
-2.78514981e-01 4.87959117e-01 -6.70552790e-01 -6.40695333e-01
-2.92653561e-01 -1.07367134e+00 8.91451120e-01 -8.22253674e-02
-2.80389965e-01 -8.74082923e-01 5.69085360e-01 7.22283646e-02
2.68562645e-01 4.05127674e-01 1.28973937e+00 -6.36582971e-01
-2.51952618e-01 -2.65690178e-01 -1.09316915e-01 -1.27591076e-03
4.68529612e-01 4.30351585e-01 -6.80918455e-01 -3.99142236e-01
-4.10706252e-01 3.01910937e-01 7.85518169e-01 4.73144323e-01
1.20310545e+00 1.06413193e-01 -7.93306589e-01 7.19299436e-01
1.36662686e+00 -2.01281030e-02 5.50487876e-01 -1.14634363e-02
8.52181792e-01 5.55763125e-01 5.78498662e-01 4.20619458e-01
-4.85435799e-02 5.90550363e-01 1.02313980e-02 6.99922293e-02
5.62934801e-02 -1.93891063e-01 3.31180811e-01 1.00331485e+00
6.35458454e-02 -7.49478489e-02 -9.38405812e-01 4.52183753e-01
-1.80931604e+00 -8.07453454e-01 -5.17642260e-01 2.18577075e+00
7.03913271e-01 -6.07718289e-01 1.75034970e-01 7.57252872e-02
9.07624662e-01 -2.37926140e-01 -5.57855248e-01 4.73340452e-02
-4.26301211e-01 -2.98147619e-01 2.85170197e-01 4.11576182e-01
-6.28576338e-01 5.53396165e-01 5.90399647e+00 8.24999452e-01
-1.10901272e+00 -6.68642595e-02 6.51991725e-01 -2.75048405e-01
-2.83893496e-01 -3.62961516e-02 -7.64480472e-01 6.57795608e-01
5.95484376e-01 -4.75651771e-01 2.57108271e-01 6.81935430e-01
6.52844071e-01 -1.02992542e-01 -1.12999070e+00 1.24042797e+00
-2.74335504e-01 -1.50416601e+00 6.40268475e-02 3.41302574e-01
8.19738626e-01 -9.29770395e-02 -2.01497413e-02 -8.32491666e-02
1.65065095e-01 -9.36765432e-01 -3.94154251e-01 3.39341223e-01
8.68130684e-01 -8.49338114e-01 9.06967163e-01 2.18458161e-01
-1.00242543e+00 5.42191379e-02 -6.79171264e-01 2.68788725e-01
1.55035570e-01 1.16663456e+00 -8.57684135e-01 3.56116295e-01
4.08381611e-01 8.21051538e-01 -1.67425618e-01 8.84447753e-01
1.74344227e-01 6.53660595e-01 -4.13544446e-01 -1.75634235e-01
-3.01658630e-01 -8.41303706e-01 7.11098909e-01 1.17955828e+00
6.39368653e-01 2.68279880e-01 2.62969602e-02 7.32955277e-01
-5.15396558e-02 3.44399273e-01 -5.88985264e-01 -3.72703701e-01
7.92903244e-01 1.54623592e+00 -1.00253427e+00 -4.60637212e-02
-1.24982275e-01 8.97926271e-01 1.81605995e-01 3.69571596e-01
-4.12811071e-01 -5.24357677e-01 1.27219546e+00 -3.41536989e-03
1.87320352e-01 -3.62653613e-01 -2.51566499e-01 -1.08945715e+00
-5.00040650e-01 -6.99783146e-01 3.75647426e-01 -2.09918261e-01
-1.24660611e+00 7.76074007e-02 -5.20154417e-01 -1.37141347e+00
1.73761472e-01 -5.51263332e-01 -3.55582237e-01 6.91384494e-01
-1.03885710e+00 -6.35341823e-01 -2.08220527e-01 1.60770789e-01
1.83139220e-01 -2.39689201e-02 1.01277876e+00 2.01763198e-01
-1.16261721e+00 2.14723438e-01 8.57374907e-01 2.51579266e-02
5.63886046e-01 -1.00283372e+00 -1.36634886e-01 5.58501184e-01
-2.59693898e-02 9.36739087e-01 6.06572807e-01 -6.21305048e-01
-1.88259804e+00 -1.28250587e+00 5.54016948e-01 -4.22047347e-01
5.52241266e-01 -6.01495922e-01 -8.24486375e-01 4.72060323e-01
-2.85598487e-01 1.17899403e-01 1.42717350e+00 7.68643543e-02
-1.89192414e-01 -2.13795856e-01 -1.03237236e+00 9.95229483e-01
8.14773440e-01 -4.88108724e-01 -1.19446274e-02 3.08441997e-01
3.99220973e-01 4.86259116e-03 -1.15380108e+00 2.03015357e-01
5.04101098e-01 -7.01979399e-01 7.62188017e-01 -5.41876793e-01
2.95686245e-01 -7.36244380e-01 -2.73302495e-01 -1.64052105e+00
-5.99783301e-01 -5.56117713e-01 2.89342552e-01 1.45353198e+00
4.23803449e-01 -6.90876663e-01 8.65804970e-01 5.42021811e-01
1.18329050e-02 -6.72537208e-01 -9.91177440e-01 -6.98293388e-01
-8.82800594e-02 -4.80264463e-02 4.90759701e-01 1.20009339e+00
6.23987257e-01 2.79084712e-01 -1.98843494e-01 -4.08201553e-02
4.99850929e-01 3.15774560e-01 1.05658305e+00 -1.25367594e+00
2.77853431e-03 -4.07284707e-01 -6.00619495e-01 -7.45464563e-01
6.30056784e-02 -1.01160955e+00 1.89670420e-03 -1.32195008e+00
5.82527697e-01 -2.20151603e-01 -2.81930834e-01 2.50384271e-01
-2.78627694e-01 2.43889093e-01 5.13833063e-03 1.56990305e-01
-3.52750212e-01 7.65427053e-01 1.12966013e+00 1.94490235e-02
-4.37395722e-01 -2.97211796e-01 -6.90840185e-01 3.73304963e-01
8.14605594e-01 -4.49393362e-01 -1.35828659e-01 1.62166461e-01
1.52853981e-01 -9.89225507e-02 -8.32601171e-03 -6.81597769e-01
2.88904816e-01 -1.76409170e-01 4.12468553e-01 -6.39692605e-01
3.71055901e-01 -7.70034254e-01 5.97410440e-01 5.98436892e-01
-2.54795372e-01 -1.11590125e-01 1.46753535e-01 7.84455061e-01
-2.75871277e-01 1.97505012e-01 7.63355374e-01 1.23322882e-01
-2.73458153e-01 2.01934740e-01 -7.02579021e-01 -1.96724921e-01
1.07923043e+00 -1.25609934e-01 -3.20914209e-01 -1.89307258e-01
-5.99460304e-01 2.95485228e-01 7.94614494e-01 1.71343740e-02
7.58746624e-01 -1.41275597e+00 -6.99184000e-01 1.92021593e-01
2.39442199e-01 -1.69591010e-01 5.09606302e-01 1.19261670e+00
-4.65512782e-01 4.54598814e-01 -4.92321774e-02 -8.49640429e-01
-1.61556911e+00 5.46378255e-01 -1.77533075e-01 1.07854083e-01
-5.89850843e-01 7.52607882e-01 2.00642988e-01 -6.47703528e-01
-3.05341482e-01 1.16664246e-01 -2.67345905e-01 1.82094306e-01
5.43571413e-01 8.08627844e-01 -2.49020845e-01 -5.10589778e-01
-3.79416764e-01 8.75851035e-01 1.81970894e-01 1.16618924e-01
1.41473043e+00 -1.87337190e-01 -7.63207138e-01 7.97856569e-01
1.48784566e+00 3.28489929e-01 -9.37178552e-01 2.18007147e-01
6.49996549e-02 -7.34286010e-01 -3.24877650e-02 -2.44731769e-01
-1.00300467e+00 8.49395692e-01 3.48558873e-01 -1.85088262e-01
1.02979147e+00 -1.64123774e-01 4.71973509e-01 3.56853127e-01
2.82647938e-01 -9.32260990e-01 -1.51719779e-01 3.64117175e-01
6.58624828e-01 -9.02575314e-01 -6.29189610e-02 -7.49532759e-01
-4.66596693e-01 9.98778224e-01 3.97936761e-01 2.84051150e-01
7.68164217e-01 2.72499412e-01 1.45900935e-01 -5.03162406e-02
-8.65770280e-01 1.33406401e-01 -2.97000527e-01 5.67555606e-01
4.94166583e-01 1.32724002e-01 -5.81869721e-01 4.95657116e-01
3.60848345e-02 6.48840442e-02 5.57695448e-01 6.29001975e-01
-1.82446361e-01 -1.18434775e+00 -3.34764034e-01 6.48752689e-01
-1.87899530e-01 -6.80849596e-04 -4.77288455e-01 5.49525559e-01
-3.48406792e-01 6.99141979e-01 4.42659967e-02 -4.66747761e-01
-8.79672915e-02 2.56360322e-01 3.18984240e-02 -5.19232690e-01
7.13119358e-02 3.25063407e-01 -3.33255231e-01 -6.58991992e-01
-1.18332058e-01 -9.11309898e-01 -1.50236070e+00 -3.15431118e-01
-3.87265474e-01 4.40010279e-01 6.83643758e-01 5.26029408e-01
1.21173692e+00 4.38716710e-01 7.53843725e-01 -4.97702599e-01
-1.64639100e-01 -7.15033114e-01 -1.05860209e+00 5.26081443e-01
-8.87258258e-03 -5.24410725e-01 -5.81328809e-01 -2.46220548e-02]
|
[6.640946865081787, 5.201248645782471]
|
14c3d7c5-f6e2-49e8-b85a-a0eca888bc90
|
hispidin-and-lepidine-e-two-natural-compounds
|
2004.08920
| null |
https://arxiv.org/abs/2004.08920v1
|
https://arxiv.org/pdf/2004.08920v1.pdf
|
Hispidin and Lepidine E: two Natural Compounds and Folic acid as Potential Inhibitors of 2019-novel coronavirus Main Protease (2019-nCoVMpro), molecular docking and SAR study
|
2019-nCoV is a novel coronavirus was isolated and identified in 2019 in Wuhan, China. On 17th February and according to world health organization, a number of 71 429 confirmed cases worldwide, among them 2162 new cases recorded in the last 24 hours. There is no drug or vaccine for human and animal coronavirus. The inhibition of 3CL hydrolase enzyme provides a promising therapeutic principle for developing treatments against CoViD-19. The 3CLpro (Mpro) known for involving in counteracting the host innate immune response. This work presents the inhibitory effect of some natural compounds against 3CL hydrolase enzyme, and explain the main interactions in inhibitor-enzyme complex. Molecular docking study carried out using Autodock Vina. By screening several molecules, we identified three candidate agents that inhibit the main protease of coronavirus. Hispidin, lepidine E, and folic acid bound tightly in the enzyme, strong hydrogen bonds have been formed (1.69-1.80&[Aring]) with the active site residues. This study provides a possible therapeutic strategy for CoViD-19.
|
['Mohamed Yousfi', 'Khedidja Benarous', 'Talia Serseg']
|
2020-04-19
| null | null | null | null |
['molecular-docking']
|
['medical']
|
[ 5.07313618e-03 -3.72814685e-01 -3.48855942e-01 2.90283393e-02
6.18200712e-02 -7.64629960e-01 1.66884974e-01 4.45299774e-01
-3.81097049e-01 1.18398154e+00 1.20077856e-01 -4.13397014e-01
2.60003597e-01 -4.59365875e-01 -4.35831070e-01 -7.92763412e-01
-2.66774327e-01 4.33834255e-01 -1.20574236e-01 2.08200160e-02
2.48106465e-01 9.92660165e-01 -7.98404336e-01 1.95934594e-01
1.39131343e+00 2.64024973e-01 5.27987123e-01 5.72619021e-01
-4.32596840e-02 3.81350458e-01 -3.83542389e-01 1.52456120e-01
8.55372399e-02 -5.98403871e-01 -1.27083331e-01 -4.88204092e-01
-1.33414626e-01 -4.10167933e-01 4.69357252e-01 5.94311059e-01
1.88778803e-01 -3.02541643e-01 1.24411869e+00 -1.02797925e+00
-4.11838859e-01 -3.07414234e-01 -4.66873765e-01 8.73752385e-02
3.30368221e-01 -4.44325842e-02 3.94336402e-01 -9.75493252e-01
1.00717974e+00 9.71510708e-01 6.70011163e-01 6.70082688e-01
-6.28764987e-01 -7.85118163e-01 -3.62087548e-01 3.95235330e-01
-1.23371804e+00 1.76029056e-01 2.26370499e-01 -7.28459299e-01
1.48515511e+00 1.69112265e-01 8.76952767e-01 4.20402884e-01
9.17806625e-01 4.38305706e-01 9.26448345e-01 2.43305862e-01
4.23522964e-02 3.17592293e-01 1.63993597e-01 4.81739223e-01
8.22665155e-01 -1.46496820e-03 -2.02745181e-02 -1.08586967e+00
3.19743633e-01 5.38234472e-01 -5.06483376e-01 -4.38563377e-01
-7.46487975e-01 1.15008783e+00 1.68395326e-01 1.48412496e-01
-6.79079175e-01 -6.85814559e-01 5.13517618e-01 -2.10096478e-01
-2.92602062e-01 3.26632351e-01 -9.45917010e-01 2.68702120e-01
-1.23440430e-01 3.74356240e-01 4.98754919e-01 4.82110173e-01
2.41050839e-01 -1.70974433e-01 1.44094154e-01 2.50632614e-01
5.40135264e-01 1.06432915e+00 -1.89161927e-01 -2.02048719e-01
-7.81856999e-02 5.91531932e-01 5.52256227e-01 -6.61765933e-01
-7.91974127e-01 1.01192053e-02 -7.12775230e-01 -7.09306225e-02
8.90129358e-02 -6.67503417e-01 -5.40549040e-01 1.55284691e+00
4.93500322e-01 3.16477120e-01 2.09919691e-01 8.02590191e-01
6.81599796e-01 1.06913829e+00 7.06449866e-01 -1.13418758e+00
1.55804813e+00 -6.72368765e-01 -8.70224714e-01 4.74275470e-01
4.37747419e-01 -8.95932913e-01 1.51077732e-01 4.19445753e-01
-5.66021502e-01 2.14038547e-02 -1.09148049e+00 6.05344653e-01
-3.00812274e-01 1.92234274e-02 5.75979531e-01 7.98218191e-01
-4.13604409e-01 2.38448247e-01 -8.79200637e-01 -8.27030897e-01
8.88255909e-02 3.60925674e-01 -3.06995779e-01 1.21941194e-01
-9.91472661e-01 1.10541642e+00 2.07411319e-01 -3.64739820e-02
-5.44139504e-01 -6.69400036e-01 -4.08742100e-01 -2.05488637e-01
-2.41036788e-01 -9.07933652e-01 6.89178884e-01 -2.40654811e-01
-1.30118620e+00 5.18161774e-01 -4.94192779e-01 -4.36813265e-01
-7.41112009e-02 -2.58572966e-01 -6.01988018e-01 3.06782097e-01
-5.04144328e-03 6.73257932e-02 1.48327023e-01 -8.68841231e-01
-4.89422202e-01 -6.57231212e-01 -5.84621906e-01 1.50311172e-01
3.64569783e-01 5.98701954e-01 2.91584820e-01 -6.67918921e-01
-2.96013802e-01 -1.04105723e+00 -3.40366244e-01 -5.29421449e-01
-1.08048813e-02 -4.95796770e-01 9.74573016e-01 -3.91584516e-01
9.91073608e-01 -1.69309390e+00 -2.41336346e-01 3.92415524e-01
-6.58805110e-03 9.68238056e-01 -2.36029662e-02 1.12931609e+00
-1.63334776e-02 3.57711166e-01 5.65688610e-02 1.03004301e+00
-4.09321815e-01 -2.97483474e-01 -3.18938106e-01 9.34230864e-01
-3.56992371e-02 6.76966131e-01 -8.54025483e-01 -1.70098543e-01
-5.57679078e-03 7.44913101e-01 -5.38542092e-01 2.73269325e-01
-3.13537717e-01 4.54121232e-01 -1.18515968e+00 6.56027377e-01
1.33036900e+00 -4.93863299e-02 7.07307756e-01 -8.71972069e-02
-3.10088992e-01 -4.72034588e-02 -5.47689259e-01 8.88147473e-01
5.07018447e-01 1.88005388e-01 -1.71872657e-02 -3.04989874e-01
5.55406809e-01 7.71235824e-01 3.90676498e-01 -4.71093655e-01
3.30974072e-01 1.51311800e-01 -7.09150508e-02 -6.08816743e-01
-2.07828850e-01 -3.03648442e-01 4.88234609e-01 1.32738709e-01
-6.03605568e-01 5.68112433e-01 2.50821531e-01 2.82581806e-01
7.35830367e-01 -1.51455313e-01 7.82382548e-01 -6.54635370e-01
9.14290488e-01 2.96115458e-01 1.03371298e+00 4.64994937e-01
-3.41855168e-01 -1.16310425e-01 3.94596457e-01 -5.62404573e-01
-8.52179587e-01 -9.76948857e-01 -5.09947002e-01 4.36577737e-01
2.69600332e-01 -4.07557130e-01 -8.60054791e-01 -1.63373023e-01
-9.03849304e-02 4.29103464e-01 -4.35362279e-01 3.30384165e-01
-5.46620309e-01 -6.16148293e-01 4.86126393e-01 5.86052760e-02
5.24198294e-01 -1.08958185e+00 -5.70320904e-01 3.47074538e-01
1.51687443e-01 -7.33268976e-01 -4.68653023e-01 -2.68547565e-01
-1.08822751e+00 -1.67201650e+00 -3.53142887e-01 -9.24318194e-01
4.04385298e-01 3.02544534e-01 2.47876659e-01 3.56227160e-01
-3.39532465e-01 -3.58159006e-01 -1.64640665e-01 -8.18865001e-01
-2.13641450e-01 -3.99456888e-01 4.50577646e-01 -6.46831393e-01
8.60769629e-01 -1.88573435e-01 -9.69106615e-01 2.07898796e-01
-2.83371508e-01 -1.04192026e-01 6.15266740e-01 5.18442333e-01
3.52331072e-01 -5.49423397e-01 7.71983206e-01 -1.02293646e+00
5.10687709e-01 -6.43321097e-01 -7.81143069e-01 2.19214559e-01
-6.59713864e-01 -3.53727311e-01 6.68952882e-01 -5.98690175e-02
-1.04110765e+00 -9.33129787e-02 -3.49508524e-02 2.24096119e-01
-3.43185626e-02 3.39183003e-01 -3.46600473e-01 2.48389825e-01
3.94034147e-01 9.36746411e-03 2.61941850e-01 -4.51159984e-01
-2.36651927e-01 6.41110778e-01 7.42557123e-02 -9.21467096e-02
6.48982167e-01 4.86537129e-01 1.92229711e-02 -1.13917267e+00
-3.88057768e-01 -7.20108509e-01 2.04686359e-01 1.50958762e-01
1.71614230e+00 -9.26004410e-01 -1.53259575e+00 5.74464619e-01
-1.36361969e+00 3.11654449e-01 9.25574005e-01 1.10060120e+00
-6.58729747e-02 5.01953304e-01 -8.72026443e-01 -5.28037250e-01
-1.06908870e+00 -8.43122721e-01 2.49972474e-02 4.74219710e-01
-4.41058502e-02 -6.93579793e-01 8.51620197e-01 3.59343410e-01
3.42019886e-01 6.42430186e-01 1.33088481e+00 -8.36185038e-01
-6.74681485e-01 -5.26073091e-02 -2.95783967e-01 7.25857262e-03
1.18636474e-01 1.68379024e-01 -4.61179852e-01 -4.78258282e-01
4.00123335e-02 1.52119115e-01 2.91243494e-01 6.93683803e-01
1.91980019e-01 -4.26305324e-01 -9.24133539e-01 5.73201418e-01
1.52261674e+00 1.12139595e+00 8.80188763e-01 3.19320738e-01
2.78684527e-01 5.25463186e-02 9.56259310e-01 5.14975548e-01
-2.27357507e-01 3.56882364e-01 2.59891540e-01 -6.78264722e-02
6.89219952e-01 -1.92793086e-02 7.61036575e-02 4.25527394e-01
-7.01191604e-01 -4.02330548e-01 -6.03268147e-01 9.11732540e-02
-1.49777877e+00 -1.25132656e+00 -7.60086000e-01 2.09849811e+00
6.15296066e-01 -2.99146593e-01 -1.01609966e-02 -8.68831515e-01
5.76806426e-01 -3.59630376e-01 -7.09263444e-01 -7.08308816e-01
-2.32941851e-01 2.07842827e-01 6.93089247e-01 5.17739177e-01
-8.77480447e-01 5.28221369e-01 6.55395508e+00 2.10042164e-01
-8.54883969e-01 -2.84992754e-01 -1.25160620e-01 2.00401440e-01
-1.78690776e-01 3.44833255e-01 -6.43258154e-01 2.42596552e-01
5.78977108e-01 -4.24571037e-01 9.25223157e-02 8.02269161e-01
6.46651745e-01 -3.23946811e-02 -4.38603610e-01 9.00413394e-01
7.66812041e-02 -1.71329677e+00 1.69167921e-01 3.92983377e-01
8.05851519e-01 1.68474227e-01 -6.24381065e-01 -2.10316833e-02
-4.27717507e-01 -7.94511557e-01 -1.43084064e-01 5.69721997e-01
2.96007425e-01 -1.12093937e+00 9.46594000e-01 2.51237929e-01
-1.12537944e+00 3.04948032e-01 -3.72813106e-01 5.42290397e-02
2.45832711e-01 1.46425758e-02 -1.21401489e+00 5.19173965e-02
3.54794711e-01 2.56130159e-01 4.06240858e-02 1.19985437e+00
-2.06969351e-01 2.43572518e-01 -1.40450940e-01 -3.60161602e-01
6.40751645e-02 -7.44656742e-01 5.14536560e-01 1.16019762e+00
-1.83477983e-01 9.07092512e-01 1.36335805e-01 3.42165798e-01
2.08211944e-01 6.55008256e-01 -5.69274783e-01 -1.37797058e-01
2.78595239e-01 9.13998067e-01 -6.20318472e-01 -4.66964155e-01
-4.43225205e-01 6.86020911e-01 -2.24058196e-01 5.23628175e-01
-8.43284011e-01 -4.60996598e-01 1.25286376e+00 1.67977571e-01
4.77024943e-01 1.18536420e-01 2.64457375e-01 -7.77427256e-01
-5.40455699e-01 -1.07954085e+00 5.76349258e-01 -5.01163483e-01
-8.73349130e-01 3.52339000e-01 -7.07546026e-02 -9.46347833e-01
1.20001927e-01 -6.03556156e-01 -7.55740285e-01 8.75425935e-01
-1.01915097e+00 -8.27850223e-01 3.61652225e-02 2.20189661e-01
2.11288750e-01 -1.44190624e-01 1.31042910e+00 -1.69937596e-01
-5.01659095e-01 7.98860937e-02 6.57403409e-01 -2.97697812e-01
8.00908625e-01 -6.06692076e-01 -6.93806559e-02 4.58592981e-01
-1.06034255e+00 1.33120847e+00 9.29906189e-01 -1.24102974e+00
-1.43883777e+00 -7.70245016e-01 1.31640959e+00 -1.81538105e-01
2.86586136e-01 -1.46590590e-01 -7.52587438e-01 5.62298894e-01
6.52515113e-01 -7.65480697e-01 1.31078994e+00 -5.37410438e-01
-3.07116568e-01 5.89832664e-01 -1.36109579e+00 6.11857474e-01
4.33282465e-01 -1.93448231e-01 -5.46388984e-01 5.35083890e-01
4.79006678e-01 -1.12848021e-01 -8.26679230e-01 7.23271787e-01
9.25722599e-01 -8.22529972e-01 6.68406487e-01 -1.00282919e+00
-4.38923597e-01 -9.91042495e-01 3.11557233e-01 -5.97368777e-01
-3.80099446e-01 -6.91194117e-01 3.18995565e-01 3.89220864e-01
3.63256454e-01 -7.84583628e-01 5.57837188e-01 4.26500618e-01
1.44507051e-01 -6.83426142e-01 -3.55527908e-01 -5.10411322e-01
-2.09312484e-01 3.78812999e-01 4.30740148e-01 8.75103533e-01
2.67392516e-01 4.73086566e-01 -4.28693354e-01 2.34336123e-01
6.53269589e-01 2.67139018e-01 6.90836549e-01 -1.29659784e+00
1.68462008e-01 1.14255346e-01 -1.28647953e-01 -6.11515224e-01
-2.05863252e-01 -6.32432640e-01 -8.17201853e-01 -1.73616457e+00
6.57506943e-01 1.05306283e-02 -2.44811946e-03 2.41063982e-01
1.77211955e-01 -1.82767555e-01 1.73559517e-01 1.07793070e-01
-2.15542793e-01 2.65625626e-01 9.49083209e-01 1.00604109e-01
-5.88335276e-01 -1.08191431e-01 -3.74935597e-01 7.31206596e-01
1.16002607e+00 -5.67235410e-01 -4.62436259e-01 4.12357569e-01
3.47222745e-01 1.13591354e-03 -1.99399352e-01 -1.24991916e-01
-8.28769058e-02 -7.23090053e-01 4.08676744e-01 -1.32305908e+00
-8.35533626e-03 -7.91853845e-01 9.47796822e-01 1.17713916e+00
7.13908851e-01 -1.83173064e-02 3.31261784e-01 4.19479579e-01
3.78239959e-01 -2.16565058e-01 7.59658635e-01 1.43951386e-01
-3.58163714e-01 1.61984429e-01 -1.27482796e+00 -9.67902839e-02
1.25264871e+00 -3.76723319e-01 -5.54070473e-01 2.14005277e-01
-5.52819788e-01 2.00678095e-01 7.69756496e-01 2.71265447e-01
5.73626101e-01 -8.80429983e-01 -7.20569193e-01 3.51187766e-01
3.01483631e-01 -7.05931008e-01 7.00539291e-01 8.99555326e-01
-1.41360939e+00 1.00089097e+00 -4.55456853e-01 -3.11545819e-01
-1.57302094e+00 1.08801472e+00 3.85593265e-01 1.49542183e-01
-3.38703811e-01 1.51214495e-01 5.86161137e-01 1.19284950e-01
4.44393419e-02 4.66972142e-02 -6.63918436e-01 -1.09606311e-01
8.71357262e-01 5.09642065e-01 -7.83499405e-02 -6.79908693e-01
-9.38715637e-01 7.43999779e-01 -4.36017305e-01 8.59336793e-01
1.24196017e+00 1.52856812e-01 -5.73669434e-01 -3.77198994e-01
1.18771374e+00 3.58842641e-01 -4.52814907e-01 2.99492180e-01
-1.61984295e-01 -2.71946609e-01 -5.61066985e-01 -8.85959566e-01
-4.24041748e-01 -1.17530279e-01 1.00468397e+00 -4.76718873e-01
8.25335503e-01 -7.01048002e-02 1.03355467e+00 3.98372293e-01
2.84886181e-01 -7.34592378e-01 -6.79262340e-01 3.30266058e-01
8.33346725e-01 -8.03661346e-01 6.61116280e-03 -5.20152628e-01
-5.59964776e-01 7.89420485e-01 4.48420227e-01 -2.13972360e-01
7.31315434e-01 -3.34016271e-02 4.18273181e-01 -4.40491110e-01
-1.01643550e+00 3.61119747e-01 -5.88335618e-02 7.52405167e-01
7.88609922e-01 2.23962903e-01 -1.47077751e+00 4.47631866e-01
4.76953208e-01 3.17251012e-02 4.20070678e-01 1.01526511e+00
-7.26389468e-01 -1.26993048e+00 -4.31778133e-01 2.32650444e-01
-6.20408297e-01 -3.52226421e-02 -4.41946179e-01 9.02119577e-01
1.47351667e-01 9.13357139e-01 -2.94671059e-01 3.03676099e-01
3.64060462e-01 -1.00226380e-01 2.25195572e-01 -7.86551088e-02
-5.05840361e-01 5.39678991e-01 1.97428111e-02 -2.18954414e-01
-5.51399648e-01 -6.05811119e-01 -1.99339962e+00 -4.26333040e-01
-4.75492120e-01 9.00112331e-01 7.69475996e-01 6.54888570e-01
3.69165868e-01 -4.16893184e-01 7.64253438e-01 -1.52797610e-01
-9.88490954e-02 -8.22338521e-01 -8.68203461e-01 1.21513881e-01
2.97939509e-01 -4.12269980e-01 -4.10038859e-01 1.92460383e-03]
|
[4.6884541511535645, 5.090554714202881]
|
8458458f-6531-4ef2-960a-e60e1f34f91d
|
learning-occupancy-for-monocular-3d-object
|
2305.15694
| null |
https://arxiv.org/abs/2305.15694v1
|
https://arxiv.org/pdf/2305.15694v1.pdf
|
Learning Occupancy for Monocular 3D Object Detection
|
Monocular 3D detection is a challenging task due to the lack of accurate 3D information. Existing approaches typically rely on geometry constraints and dense depth estimates to facilitate the learning, but often fail to fully exploit the benefits of three-dimensional feature extraction in frustum and 3D space. In this paper, we propose \textbf{OccupancyM3D}, a method of learning occupancy for monocular 3D detection. It directly learns occupancy in frustum and 3D space, leading to more discriminative and informative 3D features and representations. Specifically, by using synchronized raw sparse LiDAR point clouds, we define the space status and generate voxel-based occupancy labels. We formulate occupancy prediction as a simple classification problem and design associated occupancy losses. Resulting occupancy estimates are employed to enhance original frustum/3D features. As a result, experiments on KITTI and Waymo open datasets demonstrate that the proposed method achieves a new state of the art and surpasses other methods by a significant margin. Codes and pre-trained models will be available at \url{https://github.com/SPengLiang/OccupancyM3D}.
|
['Deng Cai', 'Boxi Wu', 'Wenxiao Wang', 'Wei Qian', 'Xiaopei Wu', 'Zheng Yang', 'Haoran Cheng', 'Junkai Xu', 'Liang Peng']
|
2023-05-25
| null | null | null | null |
['monocular-3d-object-detection']
|
['computer-vision']
|
[ 1.98772866e-02 -4.64750886e-01 -4.92386878e-01 -3.48514855e-01
-7.40973055e-01 -4.26969856e-01 5.64375937e-01 1.35383755e-01
-3.74550223e-01 8.98384750e-01 1.53176472e-01 -1.62363067e-01
-2.15730667e-01 -6.92936063e-01 -8.03357601e-01 -7.33040452e-01
-2.73234010e-01 6.71427488e-01 1.25303671e-01 2.38162994e-01
4.05285627e-01 7.87668169e-01 -1.97597730e+00 -1.96446881e-01
6.30650878e-01 1.16006792e+00 4.44140017e-01 3.75810444e-01
-9.83241871e-02 3.91135335e-01 -1.49289474e-01 5.61863124e-01
5.43373644e-01 2.26445034e-01 -3.25902849e-01 2.42589824e-02
6.37108326e-01 -6.05481207e-01 -6.54244184e-01 4.71598148e-01
6.14975810e-01 3.67847718e-02 9.85576212e-01 -1.35482836e+00
-5.10139614e-02 3.79979499e-02 -8.04471016e-01 4.34424460e-01
4.59199607e-01 -4.26836349e-02 9.22169507e-01 -1.25875747e+00
3.35831344e-01 1.07588065e+00 6.54243767e-01 1.39118552e-01
-1.34238887e+00 -8.58421028e-01 3.18549842e-01 1.47276178e-01
-2.09673524e+00 -2.94676930e-01 6.09320462e-01 -6.58238888e-01
1.02271748e+00 4.58969831e-01 8.70826244e-01 8.89468670e-01
1.08636200e-01 9.25193310e-01 1.06762993e+00 -2.32181832e-01
2.82157242e-01 -1.23349659e-01 -7.17686191e-02 6.80981815e-01
3.29196930e-01 3.44494134e-01 -7.44483232e-01 -2.38700271e-01
9.18463230e-01 4.69120175e-01 -2.28163943e-01 -9.46030319e-01
-8.39857996e-01 7.96913266e-01 6.93717301e-01 -5.63628525e-02
-2.51874149e-01 3.01980793e-01 9.51521993e-02 -2.34446675e-01
5.29756010e-01 3.65265682e-02 -2.63367504e-01 -7.70253763e-02
-1.03340781e+00 4.72196370e-01 1.62110701e-01 1.00464499e+00
1.27561200e+00 -2.52332151e-01 5.46749495e-02 7.30153084e-01
3.48554343e-01 8.65802169e-01 1.68037385e-01 -8.57362688e-01
4.13338423e-01 7.14205980e-01 1.87791333e-01 -5.43698251e-01
-4.96555686e-01 -2.97733784e-01 -7.31853187e-01 1.42944619e-01
-1.19465023e-01 4.05867249e-01 -1.07703531e+00 1.40178633e+00
7.23832726e-01 2.94622391e-01 -3.98447573e-01 8.64063382e-01
6.70327306e-01 5.16638577e-01 -3.32098752e-01 3.63819785e-02
7.91332960e-01 -4.94264245e-01 -1.45000800e-01 -3.82279605e-01
5.75786233e-01 -5.31660020e-01 8.57968628e-01 1.92995623e-01
-6.95536673e-01 -4.78586972e-01 -1.02513933e+00 -1.85935393e-01
-2.99656689e-01 7.28950254e-04 5.67163825e-01 5.17407179e-01
-7.71113694e-01 3.92423809e-01 -9.87580478e-01 -1.47655651e-01
6.88078403e-01 5.02656400e-01 -2.12350920e-01 -5.37199795e-01
-7.81119585e-01 7.13473260e-01 2.89091408e-01 -1.76928923e-01
-8.93086910e-01 -7.69964278e-01 -1.23616040e+00 -4.16490614e-01
2.77958304e-01 -6.03635430e-01 1.04669905e+00 2.80134171e-01
-6.85199976e-01 8.58510137e-01 -2.68116057e-01 -3.17643672e-01
4.01454985e-01 -4.22107011e-01 3.24038655e-01 -1.01806983e-01
4.91829187e-01 8.48850667e-01 6.90222681e-01 -1.30913258e+00
-8.86490583e-01 -7.98809767e-01 4.80278097e-02 4.59074318e-01
3.75800505e-02 -8.33719015e-01 -3.35750818e-01 -8.23982954e-02
6.14986658e-01 -9.22237337e-01 -3.34251672e-01 1.89162016e-01
-3.75867695e-01 -6.31805062e-02 7.33596921e-01 -2.38917410e-01
9.97858405e-01 -2.15427041e+00 -2.53723487e-02 3.22754294e-01
2.72146523e-01 -2.31158927e-01 2.94278026e-01 3.70804608e-01
3.66780519e-01 -3.47083881e-02 -2.81298488e-01 -6.04861498e-01
2.22744077e-01 4.90089983e-01 -1.73260823e-01 9.08252060e-01
5.89506254e-02 6.44430697e-01 -9.06819880e-01 -5.99408269e-01
1.05114853e+00 7.72162318e-01 -6.05993807e-01 6.19424097e-02
7.62505159e-02 5.51540494e-01 -5.39127767e-01 1.01821172e+00
8.62937331e-01 -8.06059465e-02 -2.75641859e-01 1.34901255e-01
-4.88867611e-01 3.64990145e-01 -1.17536902e+00 1.82951748e+00
-6.97653949e-01 3.32668424e-01 -1.06907040e-01 -7.07548261e-01
9.76711750e-01 -1.64068133e-01 9.22259808e-01 -7.05231607e-01
-1.51928272e-02 4.47967887e-01 -5.10475755e-01 -5.24118077e-03
5.55882096e-01 1.81000084e-02 -3.91255885e-01 1.42873555e-01
-3.69332314e-01 -6.43365264e-01 -1.76125258e-01 -5.67493364e-02
1.15841067e+00 3.06503057e-01 7.10546017e-01 -2.31458426e-01
3.00342858e-01 -1.09725289e-01 5.61880469e-01 7.30414689e-01
-2.28837237e-01 5.50440788e-01 -1.84492096e-01 -4.89833742e-01
-1.08668840e+00 -1.49233091e+00 -6.62419319e-01 5.48947930e-01
4.23955739e-01 -4.32179719e-01 -1.18067652e-01 -3.11760664e-01
6.86932802e-01 4.39739764e-01 -5.59270918e-01 3.73204127e-02
-4.25068825e-01 -4.60751951e-01 1.30207986e-01 6.49731696e-01
1.50269613e-01 -4.60283637e-01 -9.66130853e-01 -6.95308968e-02
-2.31427662e-02 -1.07003164e+00 -2.15964571e-01 6.10548735e-01
-8.89916301e-01 -9.27859008e-01 -5.07647157e-01 -4.59582508e-01
5.93962312e-01 7.08866537e-01 8.22512686e-01 8.25271606e-02
-9.26685154e-01 3.52226079e-01 -3.13252807e-01 -2.97755837e-01
3.85469079e-01 2.75809467e-01 3.69147837e-01 -6.28657520e-01
5.56620896e-01 -1.01309514e+00 -8.00307214e-01 3.36869717e-01
-4.95791197e-01 1.97466314e-01 5.30275464e-01 6.41760409e-01
1.20546699e+00 -4.86821793e-02 1.17505141e-01 -4.00967956e-01
-3.33975971e-01 -5.95498443e-01 -7.91965604e-01 -5.28074443e-01
-4.57195520e-01 -8.74034464e-02 1.91700801e-01 5.35425469e-02
-4.98182237e-01 5.53751588e-01 -1.09564312e-01 -7.81199455e-01
-3.17641884e-01 1.29789934e-01 -9.24217626e-02 4.07600775e-03
6.34281576e-01 4.16401893e-01 -4.53204572e-01 -5.74000180e-01
2.62525320e-01 8.20240080e-01 3.17231387e-01 -6.16762578e-01
1.09621751e+00 8.78050268e-01 1.13003612e-01 -1.01168191e+00
-1.03746092e+00 -1.04193234e+00 -9.71942961e-01 -2.05533400e-01
4.87620592e-01 -1.41882694e+00 -4.84713823e-01 7.84054026e-02
-6.45236075e-01 -3.93697470e-01 -4.13899809e-01 5.16962171e-01
-8.15144956e-01 2.31277853e-01 -1.86515555e-01 -1.31591773e+00
-1.05782345e-01 -9.00797486e-01 1.69685328e+00 -1.04561094e-02
-1.07682578e-01 -4.81732547e-01 2.67641932e-01 4.02026981e-01
-1.83289032e-02 5.59736848e-01 5.67229867e-01 1.03357926e-01
-1.06177700e+00 -2.87799418e-01 -1.94370359e-01 1.75069198e-02
2.59536237e-01 -2.91024715e-01 -1.08243263e+00 -3.62805933e-01
-2.42668077e-01 -3.93669575e-01 1.06851995e+00 6.19655132e-01
1.45774031e+00 1.37112364e-01 -7.24877357e-01 7.06585824e-01
1.49245143e+00 -9.23994705e-02 3.53802323e-01 2.48531893e-01
7.51569092e-01 2.99779594e-01 1.00262821e+00 1.02945304e+00
6.62838221e-01 8.10336888e-01 7.95863688e-01 2.38859318e-02
1.17061762e-02 -5.25062323e-01 -3.97004012e-04 5.22506595e-01
1.00737117e-01 2.13760301e-01 -1.09039307e+00 7.16867149e-01
-1.91167307e+00 -7.55644023e-01 -9.58833657e-03 2.44453049e+00
5.08950531e-01 1.49435416e-01 1.16765782e-01 3.31964552e-01
5.19414723e-01 2.46241137e-01 -8.82740557e-01 1.75325841e-01
4.42699455e-02 3.26390445e-01 8.66004884e-01 6.87935531e-01
-1.32179165e+00 7.97006488e-01 5.34080601e+00 7.55270481e-01
-7.53662229e-01 1.48078129e-02 3.21512997e-01 -6.86636329e-01
-3.17989320e-01 -1.33294597e-01 -1.20179033e+00 3.97988260e-01
5.05716085e-01 5.56931607e-02 3.67346942e-01 9.80486512e-01
2.92201251e-01 -5.02236724e-01 -1.08919156e+00 1.39409149e+00
2.02107504e-02 -1.05297196e+00 -3.04427087e-01 5.92555463e-01
6.96579993e-01 4.07664061e-01 -2.29531573e-03 4.01975155e-01
9.35129225e-02 -1.14434481e+00 8.05760443e-01 2.68415511e-01
9.44786429e-01 -7.25643694e-01 4.68939781e-01 6.59701407e-01
-1.75332844e+00 -1.83809936e-01 -6.00648522e-01 -5.05483687e-01
1.99124932e-01 7.93599784e-01 -9.77656901e-01 4.78963315e-01
8.29561889e-01 8.16381872e-01 -2.64466852e-01 1.26559186e+00
-1.39438435e-01 2.56195832e-02 -8.30880761e-01 7.84269348e-02
1.85772449e-01 3.73338759e-02 2.77653605e-01 8.47644269e-01
4.95851040e-01 8.17321520e-03 5.51129997e-01 7.94955075e-01
9.06956717e-02 -4.36607795e-03 -9.73763347e-01 3.34589154e-01
9.63972986e-01 9.65609372e-01 -4.79777098e-01 3.15538086e-02
-3.18912148e-01 7.22090900e-01 3.99955094e-01 -2.04172313e-01
-8.72393370e-01 1.55944407e-01 9.11513925e-01 5.50683260e-01
5.05483150e-01 -5.31519294e-01 -2.67654538e-01 -9.85849738e-01
1.27016261e-01 -1.15361460e-01 1.83083549e-01 -5.44027865e-01
-1.07455885e+00 1.84555069e-01 3.54749471e-01 -1.27275872e+00
-1.44047812e-01 -5.67924678e-01 -1.51371211e-02 6.17983282e-01
-1.78859282e+00 -1.04303634e+00 -6.36429965e-01 4.38924909e-01
6.14739418e-01 2.85772800e-01 7.51252115e-01 5.88834941e-01
-3.46780300e-01 1.58946022e-01 2.48294502e-01 -3.70167166e-01
2.34844044e-01 -1.23551679e+00 4.79040384e-01 2.90713161e-01
2.37910524e-01 1.24788485e-01 4.95959818e-01 -7.43383050e-01
-1.37917399e+00 -1.25182903e+00 6.18644536e-01 -6.11067951e-01
2.66278356e-01 -6.81436121e-01 -4.04050499e-01 6.07861698e-01
-6.82301521e-01 1.17607415e-01 7.13074446e-01 1.24189585e-01
-3.01041156e-01 3.06553226e-02 -1.20923805e+00 3.49084735e-01
1.70736682e+00 -5.15016854e-01 -3.42313498e-01 3.49326670e-01
5.05985558e-01 -6.15367293e-01 -7.97622681e-01 7.43583858e-01
6.49965584e-01 -9.36211705e-01 1.36084807e+00 1.27932698e-01
9.11397338e-02 -4.99565959e-01 -9.51831460e-01 -8.59062791e-01
-2.96034545e-01 -7.49249682e-02 -3.55614245e-01 7.44950891e-01
1.03993565e-01 -3.63656819e-01 1.13757265e+00 1.39680848e-01
-2.58842915e-01 -8.61311734e-01 -1.51317704e+00 -9.12816823e-01
-1.77824069e-02 -6.50095820e-01 5.56261241e-01 6.53622448e-01
-1.88631907e-01 2.12074503e-01 -3.86602372e-01 4.33950156e-01
9.36050296e-01 3.99437964e-01 1.05701470e+00 -1.41981113e+00
1.43720254e-01 -8.04702863e-02 -7.24913955e-01 -1.45661652e+00
9.06129554e-02 -9.66969192e-01 2.02241853e-01 -1.63191915e+00
2.80995458e-01 -9.32259500e-01 -2.30314776e-01 4.66973245e-01
1.10282280e-01 5.48829615e-01 -3.28758918e-02 3.93186688e-01
-5.27238131e-01 1.03967130e+00 8.97540331e-01 -1.12242706e-01
-2.15037435e-01 -9.36410427e-02 -3.33071619e-01 5.06240666e-01
1.01015568e+00 -3.70194256e-01 -3.31617266e-01 -2.74680376e-01
-2.37796828e-01 -1.89809620e-01 5.42957723e-01 -1.29534245e+00
-1.35091946e-01 -2.59178609e-01 6.78064942e-01 -1.49239993e+00
1.16092300e+00 -8.40017796e-01 1.10016830e-01 5.08167624e-01
2.87531942e-01 -2.24796876e-01 1.86568618e-01 5.74380875e-01
2.23747417e-01 4.35033254e-02 5.95549405e-01 -2.48933449e-01
-7.98573196e-01 7.73429930e-01 -6.75581470e-02 -2.31707752e-01
1.06216633e+00 -5.56786776e-01 -2.81986389e-02 -8.13449547e-02
-3.76195252e-01 4.70621198e-01 9.36587632e-01 2.65900910e-01
8.61682951e-01 -1.54267871e+00 -5.43323040e-01 2.36993298e-01
3.42140377e-01 6.17009223e-01 4.08080220e-01 7.57875383e-01
-4.06942993e-01 7.19502270e-01 -1.44691855e-01 -1.30472517e+00
-1.13524568e+00 2.84806222e-01 2.16962636e-01 8.51421282e-02
-7.89352417e-01 7.69835889e-01 3.80743235e-01 -7.23520696e-01
3.46995115e-01 -4.12041038e-01 2.51968443e-01 -1.53600693e-01
2.87588060e-01 2.47128591e-01 -8.96448940e-02 -7.52860546e-01
-6.39795005e-01 7.09638953e-01 8.68857503e-02 1.32996038e-01
1.47308373e+00 -3.92066032e-01 4.50535685e-01 6.67619348e-01
1.01536739e+00 -1.97244138e-01 -1.55454147e+00 -3.10762316e-01
-2.86388874e-01 -1.15092230e+00 2.16936201e-01 -2.84207970e-01
-5.66393912e-01 8.59203458e-01 8.44543219e-01 -1.54488474e-01
6.76145136e-01 3.85747194e-01 6.58453405e-01 2.90987045e-01
8.57924402e-01 -9.05008137e-01 4.86355573e-02 5.12223244e-01
7.32695758e-01 -1.24454343e+00 4.14295346e-01 -5.87872028e-01
-5.58505533e-04 5.49700260e-01 6.74402058e-01 -9.05943140e-02
6.97071016e-01 2.32921258e-01 -2.55364120e-01 -3.03515732e-01
-3.78576726e-01 -5.45595765e-01 -6.29572645e-02 7.02395260e-01
1.95471600e-01 2.96660095e-01 1.21675454e-01 1.34686574e-01
-4.97446001e-01 -2.15401188e-01 1.67903244e-01 1.25859904e+00
-8.51433098e-01 -9.43451285e-01 -7.05357194e-01 7.20406234e-01
3.00161719e-01 9.50113498e-03 -1.50037393e-01 9.46618676e-01
3.50413382e-01 5.67971051e-01 2.63834715e-01 -4.86771792e-01
2.81146646e-01 -1.55995607e-01 7.53521681e-01 -9.01858270e-01
3.55597198e-01 7.88177177e-02 -1.59072429e-01 -5.86443603e-01
-2.50548631e-01 -9.25315142e-01 -1.29622746e+00 -4.30112392e-01
-3.35005373e-01 -1.39131516e-01 7.53184021e-01 6.80610716e-01
2.61427283e-01 1.75608903e-01 8.11205208e-01 -1.52375758e+00
-3.46675128e-01 -6.56763971e-01 -8.59963596e-01 -1.11888992e-02
4.47743207e-01 -1.37145388e+00 -4.06881213e-01 -5.56554914e-01]
|
[7.8998003005981445, -2.604307174682617]
|
92794a8d-cf1c-44d4-863c-921846fa44ae
|
support-vector-machine-based-arrhythmia
| null | null |
http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.463.8764
|
https://pdfs.semanticscholar.org/f170/a2137b48720a5f83c55ef204419f4b7551cc.pdf
|
Support vector machine based arrhythmia classification using reduced features
|
In this paper, we proposed an algorithm for arrhythmia classification, which is associated with the reduction of feature dimensions by linear discriminant analysis (LDA) and a support vector machine (SVM) based classifier. Seventeen original input features were extracted from preprocessed signals by wavelet transform, and attempts were then made to reduce these to 4 features, the linear combination of original features, by LDA. The performance of the SVM classifier with reduced features by LDA showed higher than with that by principal component analysis (PCA) and even with original features. For a cross-validation procedure, this SVM classifier was compared with Multilayer Perceptrons (MLP) and Fuzzy Inference System (FIS) classifiers. When all classifiers used the same reduced features, the overall performance of the SVM classifier was comprehensively superior to all others. Especially, the accuracy of discrimination of normal sinus rhythm (NSR), arterial premature contraction (APC), supraventricular tachycardia (SVT), premature ventricular contraction (PVC), ventricular tachycardia (VT) and ventricular fibrillation (VF) were 99.307%, 99.274%, 99.854%, 98.344%, 99.441% and 99.883%, respectively. And, even with smaller learning data, the SVM classifier offered better performance than the MLP classifier.
|
['Sun Kook Yoo', 'Kyoung Joung Lee', 'Sung Pil Cho', 'Mi Hye Song', 'Jeon Lee']
|
2005-01-01
| null | null | null |
international-journal-of-control-automation
|
['arrhythmia-detection', 'electrocardiography-ecg']
|
['medical', 'methodology']
|
[ 4.61911932e-02 -2.12783292e-01 -7.08379820e-02 1.88614726e-02
-1.43336803e-01 -6.59819365e-01 3.09870243e-01 1.34138003e-01
-1.14304043e-01 1.13865435e+00 -2.11798280e-01 -5.49728751e-01
-2.05981910e-01 -5.47472179e-01 1.60328925e-01 -8.93270254e-01
-1.93347991e-01 3.51002187e-01 -1.66302487e-01 1.06345415e-01
3.74807745e-01 8.58261645e-01 -1.66652536e+00 4.91778702e-01
9.16617274e-01 1.01146746e+00 -3.03729117e-01 6.83776259e-01
-1.69600341e-02 6.46963477e-01 -6.78536355e-01 1.47732511e-01
-9.02178213e-02 -4.58668977e-01 -5.69933653e-01 -3.41230966e-02
-5.29891789e-01 2.27113128e-01 1.83644399e-01 6.30613148e-01
5.08980155e-01 -1.01928666e-01 1.08672118e+00 -1.39272511e+00
-5.20242155e-01 2.44279236e-01 -2.90597409e-01 2.60103762e-01
3.89561951e-01 -1.78895921e-01 3.40597421e-01 -1.13064265e+00
2.99915612e-01 8.25551033e-01 9.77998316e-01 3.12571436e-01
-1.29700875e+00 -5.50652862e-01 -5.47530770e-01 1.44335479e-01
-1.46452320e+00 -1.34616330e-01 8.59182656e-01 -8.64631355e-01
9.89544272e-01 6.78087592e-01 1.05057716e+00 7.19235182e-01
5.83766818e-01 4.19359982e-01 1.37735546e+00 -8.63597393e-01
2.33707562e-01 8.31555545e-01 6.56508148e-01 6.13737524e-01
5.07031679e-01 1.97440907e-01 -7.78880809e-03 -7.69163370e-01
4.42002416e-01 2.44170707e-03 -2.83748806e-01 -1.04744561e-01
-1.11050510e+00 8.92358899e-01 -3.01734865e-01 7.80563891e-01
-5.73980689e-01 -8.93835604e-01 6.26233757e-01 2.93577790e-01
2.50078648e-01 4.86728281e-01 -6.71846092e-01 -2.35375121e-01
-8.14644814e-01 -2.30507955e-01 9.86469269e-01 3.72858256e-01
3.68085027e-01 5.34686208e-01 -5.70021905e-02 8.93683195e-01
1.74755320e-01 6.88832998e-01 8.57274294e-01 -6.23733222e-01
3.64746936e-02 7.03100264e-01 -8.68122503e-02 -1.02741098e+00
-6.11375451e-01 -3.70069236e-01 -1.26866674e+00 2.75117010e-01
3.84973675e-01 -3.44294757e-01 -6.43309772e-01 1.10242391e+00
-2.40989372e-01 -9.31309722e-03 6.31252110e-01 5.79244554e-01
8.22617650e-01 6.59285545e-01 1.16503514e-01 -8.27116191e-01
1.20027721e+00 -4.58970368e-01 -8.60325038e-01 1.76715940e-01
6.11573577e-01 -8.96085322e-01 7.38580167e-01 7.57817447e-01
-6.04135334e-01 -7.04220176e-01 -1.03694618e+00 7.81604707e-01
-2.19064578e-02 6.59104049e-01 6.38728678e-01 9.92021739e-01
-7.18193352e-01 6.83613956e-01 -7.08147407e-01 -3.45128328e-01
2.42811367e-01 4.23486143e-01 -4.85029578e-01 4.56245631e-01
-1.00887096e+00 1.06148136e+00 2.92956710e-01 -6.77407021e-04
2.23874189e-02 -4.43242282e-01 -7.10797429e-01 4.07359861e-02
-6.12050354e-01 -3.92117232e-01 5.30019820e-01 -8.74621570e-01
-1.53280783e+00 6.03017986e-01 -3.89393449e-01 -2.35362098e-01
-8.95288140e-02 1.93465844e-01 -8.07332754e-01 3.43913585e-01
-5.78132309e-02 -7.58171529e-02 8.30731452e-01 -1.12151623e+00
-5.36585331e-01 -3.86767894e-01 -7.29220629e-01 -1.84197038e-01
-2.57828802e-01 1.17046893e-01 6.81441724e-01 -5.01314640e-01
4.82923746e-01 -1.03167570e+00 9.30285677e-02 -4.90486681e-01
-2.27858111e-01 -4.04904962e-01 8.24077725e-01 -9.54861522e-01
1.36427343e+00 -2.36980987e+00 5.07428013e-02 5.79635680e-01
-1.41374737e-01 7.54694521e-01 4.93956268e-01 4.29554224e-01
-3.97214979e-01 1.26987413e-01 -3.79840434e-01 3.52420658e-01
-4.71596450e-01 2.10332736e-01 -2.71450758e-01 4.04029191e-01
2.58334607e-01 3.90319645e-01 -4.58264887e-01 -5.77095747e-01
3.64580125e-01 5.53467274e-01 -1.75169617e-01 -1.83666125e-01
9.83524621e-01 3.46295506e-01 -3.27004701e-01 9.54296589e-01
5.43644845e-01 1.60968855e-01 1.36182532e-01 -2.61160791e-01
-1.91165105e-01 -4.45059270e-01 -1.09060729e+00 8.32416415e-01
-2.11601436e-01 7.25346446e-01 -3.82262766e-01 -1.34398448e+00
1.77131867e+00 8.76684725e-01 7.26263046e-01 -1.33166283e-01
1.84050929e-02 5.26522875e-01 2.92461663e-01 -7.65167832e-01
-2.74675280e-01 -3.15793961e-01 1.28955498e-01 1.71874225e-01
1.95305399e-03 2.36863550e-02 -1.37977883e-01 -3.33316594e-01
6.01112604e-01 -2.69186318e-01 8.42370212e-01 -5.98595917e-01
1.04423618e+00 1.61855757e-01 6.65238976e-01 2.96519548e-01
-3.49065542e-01 4.51960653e-01 6.39936268e-01 -7.97293365e-01
-6.98629379e-01 -8.50347996e-01 -7.28791416e-01 -2.78654788e-03
-2.48799667e-01 -2.16813922e-01 -4.30478334e-01 -4.77939129e-01
1.32163063e-01 7.42697656e-01 -1.33999035e-01 -2.47424185e-01
-4.00650889e-01 -1.19492638e+00 5.44680417e-01 4.19635683e-01
3.80121291e-01 -1.20224881e+00 -5.40515721e-01 2.57464588e-01
-8.82894546e-03 -6.33290291e-01 4.23633546e-01 2.93920904e-01
-1.17382348e+00 -1.17344999e+00 -5.14478266e-01 -8.98613513e-01
5.03729641e-01 -2.32616559e-01 6.20196760e-01 1.62547395e-01
-3.93903315e-01 -3.70735824e-02 -4.37671959e-01 -5.40688396e-01
-7.00509310e-01 -2.99518049e-01 6.13710344e-01 -9.47796553e-03
4.59689349e-01 -6.45649791e-01 -8.39812830e-02 2.59033322e-01
-2.68303692e-01 -4.00538713e-01 5.86920440e-01 1.06131375e+00
4.14154232e-01 3.23969990e-01 1.00620341e+00 -6.17119133e-01
6.99254036e-01 -2.89822012e-01 -9.28358212e-02 2.46813241e-02
-9.91290271e-01 -3.56391609e-01 9.17382240e-01 -4.54060167e-01
-7.76294053e-01 2.67279476e-01 -1.19394332e-01 -3.83997977e-01
-4.70212787e-01 4.41876888e-01 3.09062637e-02 -1.66048616e-01
1.00902975e+00 6.28017187e-01 4.65981424e-01 -4.27227855e-01
-4.60791439e-01 1.27215564e+00 2.16419131e-01 -1.69564754e-01
6.22208953e-01 1.85038373e-01 1.04337744e-01 -1.12248862e+00
-3.45272094e-01 -2.87059754e-01 -9.15831387e-01 -1.52221829e-01
7.51624465e-01 -7.15836346e-01 -5.99060059e-01 5.63842535e-01
-9.39874709e-01 4.01077390e-01 -2.38901868e-01 1.11394942e+00
-4.39687520e-01 7.27352858e-01 -4.47686493e-01 -1.29501760e+00
-7.39714146e-01 -8.76166046e-01 2.07212180e-01 7.87982121e-02
-7.36157000e-01 -7.76617825e-01 -7.19463751e-02 1.05373643e-01
3.68547469e-01 5.92863262e-01 1.16711676e+00 -1.04047406e+00
6.21382058e-01 -6.43577754e-01 8.02454054e-02 8.63855660e-01
5.44730663e-01 4.48978841e-01 -1.01354384e+00 -1.56539634e-01
5.29162109e-01 3.13975602e-01 4.59547997e-01 4.92180765e-01
8.44319880e-01 -2.42954671e-01 -4.59718883e-01 5.49453437e-01
1.08932412e+00 1.17924476e+00 6.29743695e-01 4.22702700e-01
3.26486468e-01 2.49511376e-01 6.38359249e-01 4.18887466e-01
-1.56710923e-01 4.31892812e-01 -9.46717709e-02 -1.75669149e-01
2.09051877e-01 3.40304852e-01 2.64412820e-01 7.67856777e-01
-5.66537321e-01 4.10489827e-01 -1.04682338e+00 1.44172862e-01
-1.47887456e+00 -1.14664435e+00 -6.53503418e-01 2.09574604e+00
5.79993427e-01 1.79307133e-01 2.53679723e-01 1.28229129e+00
8.01029265e-01 -4.21625078e-01 -8.66171047e-02 -1.10129595e+00
-3.26229423e-01 3.13660651e-01 6.22831460e-04 2.40965456e-01
-1.25834095e+00 2.01790497e-01 6.28703165e+00 4.72861558e-01
-1.29336441e+00 -3.52632970e-01 4.64486450e-01 4.02956545e-01
2.58637190e-01 3.14015634e-02 -4.48519945e-01 7.52616584e-01
9.09774482e-01 -3.07889044e-01 2.44717345e-01 9.58813131e-01
6.78326264e-02 -2.14794371e-02 -6.67954147e-01 1.30660868e+00
2.72164005e-03 -1.15111434e+00 -9.75692719e-02 -1.66561767e-01
2.87163824e-01 -6.22658014e-01 -2.85290778e-01 5.13591707e-01
-8.14209998e-01 -6.89889431e-01 1.94809675e-01 6.89769864e-01
6.14042401e-01 -8.53667140e-01 1.20524609e+00 5.01228094e-01
-9.52741921e-01 -3.83548051e-01 -3.50684524e-01 -5.06649613e-01
-1.91725165e-01 8.76876891e-01 -6.98351800e-01 5.77369630e-01
6.92606986e-01 7.54696369e-01 -3.32728058e-01 8.70210230e-01
6.38840348e-02 7.91308343e-01 -2.98497766e-01 -1.23107180e-01
-2.75747359e-01 -3.53982627e-01 6.55159831e-01 1.09072196e+00
3.83092254e-01 2.55702227e-01 -5.02732880e-02 4.55919653e-01
8.87400866e-01 4.25202459e-01 -4.19552296e-01 6.42933697e-02
5.87288678e-01 1.18210709e+00 -7.85988927e-01 -5.25218308e-01
-2.51285851e-01 5.36970854e-01 -4.42633003e-01 2.66451627e-01
-6.47219539e-01 -9.22031224e-01 3.23149592e-01 -5.02800345e-02
-5.31679876e-02 2.02551991e-01 -8.40459347e-01 -1.18177891e+00
2.09331363e-01 -6.37053072e-01 4.72738832e-01 -4.91841227e-01
-1.18300903e+00 9.46995080e-01 -3.90510373e-02 -1.66357040e+00
-3.17803949e-01 -5.25178850e-01 -8.29678357e-01 1.33225644e+00
-1.01671040e+00 -6.62273884e-01 -5.15356958e-02 4.30464923e-01
1.69735804e-01 -9.93230939e-01 1.42992461e+00 8.60194713e-02
-5.93014240e-01 3.45720321e-01 1.83940679e-01 -2.83928644e-02
3.86431634e-01 -1.07780075e+00 -4.09494370e-01 5.45745552e-01
-2.24504113e-01 4.06045288e-01 4.72903520e-01 -4.24909949e-01
-1.10745323e+00 -7.75883853e-01 1.37810695e+00 -1.52560130e-01
1.71824142e-01 2.21001580e-01 -9.44937587e-01 2.11660713e-01
-1.84650987e-01 -2.19898708e-02 1.09034526e+00 -3.31894867e-02
2.19646156e-01 -3.91944982e-02 -1.51910448e+00 1.60860434e-01
1.56664848e-01 -2.23265365e-01 -9.22682166e-01 2.52295673e-01
-1.13549277e-01 -3.51324603e-02 -1.41939056e+00 4.91465092e-01
8.82500589e-01 -9.58896577e-01 1.07314134e+00 -5.47281206e-01
6.12433776e-02 -2.75675297e-01 3.09044626e-02 -9.99750376e-01
-7.77490079e-01 -3.32876474e-01 -1.44081607e-01 1.08599997e+00
4.77985799e-01 -1.09150422e+00 5.91640532e-01 4.15597260e-01
-2.44621366e-01 -7.42334843e-01 -9.79432881e-01 -9.35623586e-01
-5.67734241e-02 -2.98339408e-02 2.97335863e-01 1.22085738e+00
6.36216760e-01 2.53664851e-01 -3.15228403e-01 3.94126400e-02
4.67510313e-01 3.75414968e-01 2.15998590e-01 -1.86998224e+00
-1.25774860e-01 -2.99628556e-01 -7.85025835e-01 3.83417368e-01
1.12282194e-01 -1.02773523e+00 -7.29936242e-01 -1.44106567e+00
-1.12336047e-01 -5.28392076e-01 -5.98446846e-01 7.48673022e-01
-5.74358441e-02 1.84381470e-01 8.29971656e-02 6.36199176e-01
7.48324454e-01 2.23491471e-02 9.57940400e-01 1.96366698e-01
-7.99031734e-01 7.29662478e-01 -6.44260585e-01 8.75245810e-01
1.22336555e+00 -5.55236161e-01 -1.99118257e-01 3.36016566e-01
-4.90459025e-01 4.22908187e-01 1.35829717e-01 -1.13338017e+00
-2.35853016e-01 5.20555228e-02 8.96118879e-01 -4.39130455e-01
3.09717238e-01 -6.66433156e-01 4.95741934e-01 1.14987290e+00
8.99315476e-02 -2.28590518e-01 3.77867460e-01 7.96383470e-02
-3.05216759e-01 -2.92015791e-01 6.94244206e-01 4.57902402e-01
-4.92515713e-01 -3.01812142e-01 -1.03205585e+00 -5.58027685e-01
1.30068111e+00 -7.91657627e-01 1.17507048e-01 3.90767008e-02
-1.37204027e+00 -5.17091334e-01 -2.12029554e-02 1.00636363e-01
7.84691632e-01 -1.30269849e+00 -7.63306201e-01 5.86531401e-01
-3.28751393e-02 -5.65008163e-01 1.36841908e-01 1.27309775e+00
-6.09009385e-01 5.25937438e-01 -6.97482407e-01 -6.41025186e-01
-1.59823215e+00 5.76128125e-01 7.80652761e-02 1.98446244e-01
-7.95420766e-01 3.48302901e-01 -4.80536312e-01 -8.26289877e-02
-8.94693509e-02 -3.88606548e-01 -9.40598369e-01 2.17082962e-01
2.98070341e-01 7.06123173e-01 8.36933553e-02 -6.39774501e-01
-7.61249065e-01 8.57047796e-01 3.80757809e-01 4.56956685e-01
1.13877547e+00 4.22307432e-01 -3.34687620e-01 4.63641614e-01
1.06743383e+00 -2.59679537e-02 -2.63310879e-01 1.59198537e-01
-4.27825609e-03 -2.17502430e-01 -8.57480913e-02 -8.04503679e-01
-8.02329838e-01 6.89289212e-01 7.57403672e-01 5.45755446e-01
1.31592345e+00 -3.70941281e-01 2.39458025e-01 3.93211752e-01
1.98828295e-01 -7.97040284e-01 -7.82863319e-01 2.46815771e-01
7.94752598e-01 -9.15662587e-01 -2.06010435e-02 -6.64982617e-01
-8.82555127e-01 1.87540257e+00 -7.56005272e-02 -4.15436208e-01
1.12775469e+00 2.30704829e-01 6.89547658e-02 1.77194968e-01
-6.41210914e-01 3.61737877e-01 3.61466378e-01 7.45709240e-01
3.68683785e-01 2.78436601e-01 -1.00069225e+00 1.23274493e+00
-1.93597600e-01 3.53163630e-01 5.89046836e-01 9.44695950e-01
-5.75570166e-01 -7.57799327e-01 -6.57008350e-01 8.33705485e-01
-4.25317198e-01 1.06741711e-01 -1.48098081e-01 9.81494248e-01
1.92256823e-01 1.08177459e+00 -4.45212871e-02 -5.98014951e-01
3.67912471e-01 5.22106826e-01 1.05006360e-01 -3.56841177e-01
-5.01836598e-01 1.55134559e-01 1.27247110e-01 -3.55733186e-02
-2.87616134e-01 -9.32284117e-01 -1.30622780e+00 2.35437900e-02
-2.50269443e-01 7.43366063e-01 7.45643079e-01 8.01217973e-01
1.35510251e-01 3.84011865e-01 9.47486103e-01 -4.23860908e-01
-5.15563786e-01 -9.43883836e-01 -1.09544170e+00 1.10084191e-01
1.94113374e-01 -7.03978598e-01 -9.47025239e-01 3.18561137e-01]
|
[14.137404441833496, 3.2195916175842285]
|
0a486359-5d28-43e3-97c6-f6eef158bea3
|
deep-bayesian-experimental-design-for-quantum
|
2306.14510
| null |
https://arxiv.org/abs/2306.14510v1
|
https://arxiv.org/pdf/2306.14510v1.pdf
|
Deep Bayesian Experimental Design for Quantum Many-Body Systems
|
Bayesian experimental design is a technique that allows to efficiently select measurements to characterize a physical system by maximizing the expected information gain. Recent developments in deep neural networks and normalizing flows allow for a more efficient approximation of the posterior and thus the extension of this technique to complex high-dimensional situations. In this paper, we show how this approach holds promise for adaptive measurement strategies to characterize present-day quantum technology platforms. In particular, we focus on arrays of coupled cavities and qubit arrays. Both represent model systems of high relevance for modern applications, like quantum simulations and computing, and both have been realized in platforms where measurement and control can be exploited to characterize and counteract unavoidable disorder. Thus, they represent ideal targets for applications of Bayesian experimental design.
|
['Florian Marquardt', 'Leopoldo Sarra']
|
2023-06-26
| null | null | null | null |
['experimental-design']
|
['methodology']
|
[ 2.25195870e-01 -2.94159025e-01 4.26265001e-02 -3.35114717e-01
-5.70458233e-01 -5.56029856e-01 7.27016211e-01 -3.48240579e-03
-7.86612928e-01 9.89836812e-01 -1.86783001e-01 -4.25052255e-01
-2.78698713e-01 -8.53315651e-01 -3.96969527e-01 -1.21640980e+00
-1.56561315e-01 6.99359775e-01 -1.65665850e-01 -2.13691920e-01
6.36113822e-01 1.08833694e+00 -1.37667525e+00 -1.86425969e-01
3.50936234e-01 8.94368231e-01 5.18494248e-02 6.70349121e-01
2.42354318e-01 1.37549385e-01 -4.38151419e-01 3.66897695e-02
2.03546509e-01 -4.05563623e-01 -2.91791886e-01 -6.36087000e-01
1.94287032e-01 -3.53476703e-01 -6.64829791e-01 1.47365248e+00
9.41681325e-01 1.53013870e-01 7.33227491e-01 -5.58611393e-01
-4.25739177e-02 9.57642734e-01 1.24589942e-01 3.47479224e-01
-3.05457655e-02 5.31267047e-01 9.67471421e-01 -4.48507637e-01
5.64722538e-01 1.03896165e+00 2.41316885e-01 6.59489572e-01
-1.83388448e+00 -6.47830009e-01 -8.55847836e-01 3.38110685e-01
-1.41726089e+00 -8.85415077e-01 6.98832035e-01 -2.87138164e-01
1.10326028e+00 9.76311117e-02 6.14698172e-01 1.15454555e+00
6.99709117e-01 3.09790462e-01 1.20896649e+00 -6.21625483e-01
7.69231379e-01 3.30361314e-02 1.99806914e-01 3.62268597e-01
4.46596891e-01 7.45721519e-01 -5.60543716e-01 -1.79392174e-01
5.34388006e-01 -5.41304708e-01 -2.64305025e-01 -4.70531076e-01
-1.27179837e+00 9.01858747e-01 3.28321517e-01 6.00839734e-01
-5.79964995e-01 7.78250754e-01 3.63304287e-01 3.31487149e-01
-2.62065768e-01 1.04740071e+00 6.69954568e-02 -1.12514697e-01
-9.78545010e-01 4.01516080e-01 7.57730186e-01 5.29856503e-01
7.48210371e-01 1.31683871e-01 -3.96051586e-01 8.61812476e-03
1.35184467e-01 1.06319964e+00 1.68498814e-01 -1.15618002e+00
-3.09545659e-02 -1.71162367e-01 5.14128089e-01 -5.06802857e-01
-8.45172822e-01 -4.54017431e-01 -1.09497130e+00 1.80842340e-01
2.67508596e-01 7.28215929e-03 -7.63528585e-01 1.57687140e+00
2.05633566e-01 -1.24873526e-01 2.21411020e-01 7.59386718e-01
4.80530821e-02 4.78425443e-01 -3.67298126e-01 -1.64336696e-01
1.07254910e+00 -1.18196383e-01 -8.47173214e-01 -3.83656612e-03
6.31701171e-01 -5.57106495e-01 5.51111937e-01 6.48043752e-01
-7.67860174e-01 -2.27397323e-01 -1.39394939e+00 2.12693408e-01
-3.17610055e-01 6.84700683e-02 6.71516478e-01 1.26531219e+00
-9.67699289e-01 1.11968648e+00 -1.04930520e+00 3.73609923e-03
2.13974684e-01 7.07660317e-01 -1.15395568e-01 -2.44537257e-02
-1.17431438e+00 9.76398349e-01 6.26015902e-01 6.44532382e-01
-1.05769217e+00 -2.71037877e-01 -4.58672941e-01 1.46049112e-01
-1.26617551e-01 -5.33973336e-01 1.16408503e+00 1.38779894e-01
-2.01265025e+00 5.12082696e-01 5.12649119e-02 -7.76119947e-01
-5.41314818e-02 1.72660545e-01 -1.77770063e-01 1.59789756e-01
-2.34680325e-01 2.03808963e-01 6.31564200e-01 -4.64498371e-01
-4.66922717e-03 -4.38500375e-01 -1.52423695e-01 -3.75263304e-01
-2.23143741e-01 -4.23391312e-02 4.62056935e-01 4.62458760e-01
3.03241879e-01 -1.06320310e+00 -5.08091927e-01 -4.33994621e-01
-5.51703930e-01 9.42377821e-02 2.97317773e-01 -7.68270995e-03
8.55128825e-01 -1.86555040e+00 4.49655712e-01 4.55751061e-01
1.47038132e-01 3.40271711e-01 -6.74852580e-02 7.00613141e-01
2.06170946e-01 -6.99329302e-02 -1.08161412e-01 -1.23084128e-01
3.84396672e-01 -1.23066485e-01 -9.59957018e-02 9.86366630e-01
1.67678788e-01 7.48930693e-01 -7.23483503e-01 4.81792018e-02
5.94103694e-01 2.66878217e-01 -5.41721523e-01 3.30042839e-01
-3.73894513e-01 8.97079349e-01 -6.69069707e-01 3.29821795e-01
6.62592649e-01 -7.42230415e-02 4.22875047e-01 -2.95658141e-01
-3.82100880e-01 5.66322803e-01 -1.25860739e+00 1.45352912e+00
-3.94225836e-01 6.56512856e-01 3.75213891e-01 -1.20529425e+00
9.43569541e-01 8.92672464e-02 2.96206236e-01 -9.74778354e-01
6.30309224e-01 4.98920888e-01 5.31645358e-01 -4.60095406e-01
6.29321635e-01 -4.95630801e-01 -3.01118851e-01 4.18394893e-01
2.04864442e-01 -7.00982690e-01 1.95520863e-01 1.18080474e-01
1.13820469e+00 -1.62385672e-01 2.80583858e-01 -7.27189720e-01
4.98990864e-01 -5.01621366e-01 2.67992020e-01 1.31611955e+00
-3.41935396e-01 7.82248825e-02 5.03086388e-01 -2.84265995e-01
-1.42532182e+00 -1.10760069e+00 -5.07939577e-01 3.08433115e-01
7.61048719e-02 -5.19756973e-02 -4.90376532e-01 2.39904851e-01
-9.79739800e-02 8.95714700e-01 -2.81482607e-01 -3.91242623e-01
-6.20843470e-01 -9.49663639e-01 5.62573075e-01 2.04554364e-01
1.32170185e-01 -9.29266810e-01 -3.79988939e-01 4.19008762e-01
2.19772592e-01 -1.11914480e+00 3.11351687e-01 6.66784465e-01
-5.55061877e-01 -8.45482767e-01 -1.81216240e-01 7.63679743e-02
1.38595894e-01 -2.47238472e-01 1.07815385e+00 -3.98137182e-01
-5.19295692e-01 4.23828155e-01 1.13927228e-02 -6.81200549e-02
-7.39140272e-01 1.00178584e-01 6.72980785e-01 -1.27559707e-01
5.04307151e-01 -7.64314294e-01 -6.96035147e-01 -6.91867992e-02
-7.72217572e-01 -4.89578038e-01 4.90599364e-01 1.03944969e+00
3.90869468e-01 -5.45381233e-02 2.47449368e-01 -5.45056641e-01
7.21689224e-01 -2.69002706e-01 -1.43147457e+00 1.16273649e-01
-3.33964765e-01 6.04875743e-01 7.90309906e-01 5.65393036e-03
-8.44492674e-01 -4.86480705e-02 -4.41812515e-01 -1.24281179e-02
-1.62850752e-01 4.47588503e-01 -4.51560840e-02 -4.43632454e-01
8.17116737e-01 1.51766300e-01 -2.16109663e-01 -7.46417046e-02
4.31136876e-01 7.78942049e-01 1.53806388e-01 -9.68965173e-01
6.34089828e-01 5.04232705e-01 8.56399477e-01 -9.83133495e-01
-6.19395494e-01 -2.22919241e-01 -6.24471545e-01 3.41668278e-02
7.95872450e-01 -4.94040161e-01 -1.34104526e+00 3.50468159e-01
-1.03497064e+00 -2.87774861e-01 -1.92688584e-01 1.02904177e+00
-7.03586280e-01 1.02447525e-01 -5.50398827e-01 -1.32940507e+00
-3.43113750e-01 -1.50655591e+00 1.12615001e+00 4.77300197e-01
1.97685033e-01 -8.24582279e-01 2.17306510e-01 -1.51836649e-01
8.18896294e-01 -9.35357139e-02 7.52178729e-01 -4.58386928e-01
-9.85219479e-01 -4.90592182e-01 1.38030037e-01 3.85815501e-01
-4.59088683e-01 -7.79988095e-02 -1.20108354e+00 -6.31814718e-01
1.19666874e-01 -5.56991696e-01 8.89892757e-01 5.06095707e-01
1.01770115e+00 5.89809537e-01 -3.72441471e-01 5.53272367e-01
1.27132201e+00 9.80734751e-02 6.13683939e-01 -9.91918817e-02
2.71963567e-01 3.38842571e-01 1.96117550e-01 4.76074219e-01
-3.61493349e-01 8.95617187e-01 3.85022312e-01 6.54070139e-01
2.94445902e-01 1.83307782e-01 9.86034721e-02 1.06526244e+00
2.77606070e-01 -3.91217738e-01 -8.42661977e-01 2.33531877e-01
-1.42630601e+00 -9.34432387e-01 -2.19395787e-01 2.58778214e+00
4.86065149e-01 1.59505382e-01 -3.84166181e-01 -1.87837631e-01
8.11844468e-01 4.95675653e-02 -5.54232061e-01 -4.15356487e-01
-2.11176246e-01 8.85087669e-01 6.39726818e-01 4.05676633e-01
-8.79434586e-01 6.87242687e-01 7.31122351e+00 8.81169260e-01
-1.13510668e+00 5.11350296e-02 2.94875324e-01 -1.92013830e-01
-1.87339321e-01 2.65932500e-01 -1.21296430e+00 4.50457156e-01
1.55710602e+00 7.96932504e-02 7.64948189e-01 2.85696179e-01
1.86963320e-01 -2.88664401e-01 -1.31370246e+00 8.56493056e-01
-5.24183810e-01 -1.53573120e+00 -3.85815293e-01 2.21606925e-01
5.65189183e-01 4.33013499e-01 1.99773982e-01 2.60405689e-01
8.87987539e-02 -9.10605490e-01 3.16447914e-01 6.81788206e-01
7.53043234e-01 -5.37027776e-01 9.14405346e-01 2.35031441e-01
-5.48078716e-01 -1.56647503e-01 -7.89153337e-01 -1.90927446e-01
4.79427963e-01 1.01438105e+00 -6.70727134e-01 1.41094103e-01
3.45979184e-01 4.09355834e-02 5.85375167e-02 9.92852449e-01
-1.69173300e-01 5.90007305e-01 -7.49763250e-01 -8.44485462e-01
3.09585005e-01 -6.61168158e-01 7.18356729e-01 8.64242554e-01
7.17113018e-01 -2.86978126e-01 -1.96970373e-01 1.24211788e+00
-7.90286586e-02 -3.51579905e-01 -7.10249245e-01 -5.51480412e-01
7.29864657e-01 1.20034361e+00 -6.82339549e-01 -3.68402787e-02
2.46142931e-02 3.50721329e-01 3.96064609e-01 9.38784853e-02
-4.81216520e-01 -6.21087790e-01 7.81047463e-01 -1.99208006e-01
4.49198574e-01 -7.66640186e-01 6.47336096e-02 -1.37813485e+00
-1.89759865e-01 -5.68089724e-01 -2.72736967e-01 -3.41393501e-01
-1.09733248e+00 2.18494877e-01 -1.90595523e-01 -7.07055509e-01
-3.32893133e-01 -1.14749706e+00 -4.29756939e-01 9.25371110e-01
-1.35267818e+00 -4.11036551e-01 2.11336374e-01 1.02929324e-01
-5.12360513e-01 -1.99654132e-01 1.22833455e+00 2.59112626e-01
-6.03803515e-01 2.28202686e-01 1.11930227e+00 -1.41378954e-01
4.06870335e-01 -1.38170528e+00 4.35310930e-01 9.25517321e-01
2.06990778e-01 8.71191680e-01 1.16025376e+00 -2.56466508e-01
-2.02396393e+00 -3.71046305e-01 1.45720750e-01 -1.48468629e-01
9.61979210e-01 -8.19307864e-01 -4.16631550e-01 2.64396369e-01
2.29473859e-01 3.25181819e-02 6.02130949e-01 3.92117709e-01
-2.46505246e-01 -2.26401672e-01 -1.23996472e+00 4.99036759e-01
5.24864376e-01 -9.16439176e-01 -5.66334315e-02 5.32058656e-01
2.61716902e-01 -3.43223065e-01 -9.47746933e-01 2.56485850e-01
7.64725864e-01 -1.29758883e+00 7.10515261e-01 -3.69785041e-01
-1.45748079e-01 -2.88888395e-01 -4.46281284e-01 -1.45580184e+00
-3.17574978e-01 -1.07976592e+00 -8.71595219e-02 6.40234649e-01
2.71731496e-01 -9.22407329e-01 9.32900429e-01 3.63767952e-01
-8.54801014e-02 -1.12698063e-01 -1.51070881e+00 -7.15015411e-01
3.81765991e-01 -3.63337487e-01 4.52903479e-01 4.02628124e-01
-8.44212323e-02 2.08065644e-01 -4.74891216e-01 3.13028723e-01
9.42146778e-01 1.22306198e-01 5.04998863e-01 -9.74569857e-01
-4.93741632e-01 -7.29421139e-01 -9.02282000e-01 -1.05985260e+00
2.65928537e-01 -8.27102840e-01 2.90886223e-01 -7.25263417e-01
-9.23179742e-03 -4.71738458e-01 -4.51431572e-01 -5.30496716e-01
2.63863981e-01 2.80241013e-01 -2.38800660e-01 -1.76332623e-01
-3.88913572e-01 9.14366364e-01 8.50796282e-01 -1.78371385e-01
3.87886055e-02 5.99527434e-02 -2.25189820e-01 2.58345474e-02
1.06931376e+00 -6.40937150e-01 7.67572001e-02 -2.51021922e-01
7.31222332e-01 8.12324584e-02 2.73687899e-01 -1.27095377e+00
4.11006540e-01 -6.64725080e-02 7.68262371e-02 -4.05585468e-01
7.22904682e-01 -4.15626526e-01 4.10269201e-02 5.79653859e-01
-2.49623656e-01 -4.25162077e-01 -1.11259259e-01 6.20762646e-01
-6.76137134e-02 -8.47663283e-01 1.05035090e+00 -8.51045400e-02
-2.50100404e-01 1.28990576e-01 -6.40185893e-01 -3.40896308e-01
6.97237253e-01 4.96382952e-01 -4.76862460e-01 -2.51709998e-01
-8.14300537e-01 -4.87493463e-02 2.32212290e-01 -3.75519544e-01
2.99525768e-01 -1.21382809e+00 -5.87364554e-01 2.79738605e-01
-2.30608974e-02 -4.79167134e-01 3.47644210e-01 9.36093628e-01
-7.41503835e-01 7.91562259e-01 -2.44743481e-01 -8.08617175e-01
-5.68430603e-01 3.78792375e-01 7.00711966e-01 -2.64080435e-01
-1.82755783e-01 7.39320278e-01 -3.43172431e-01 -5.84348977e-01
-2.74697691e-01 -2.07327485e-01 2.32557356e-01 -3.78916651e-01
6.75100267e-01 7.63798505e-02 2.53713280e-01 -2.85388798e-01
-2.10825741e-01 2.18988508e-01 -5.31676924e-03 -1.42626762e-01
1.30096972e+00 3.40823084e-02 -5.05301535e-01 5.45183718e-01
1.15295959e+00 -9.42081064e-02 -8.06056917e-01 -2.24742144e-01
9.91795585e-02 -2.13825077e-01 4.22790706e-01 -3.04031283e-01
-5.48829317e-01 1.40632010e+00 7.92653620e-01 5.43820918e-01
6.44471765e-01 -1.10405892e-01 3.39783788e-01 1.25214994e+00
8.55156839e-01 -9.81911242e-01 -3.38770270e-01 5.36251128e-01
3.25694919e-01 -9.96016204e-01 8.81970301e-02 1.52305916e-01
4.09625739e-01 1.40144956e+00 7.94293582e-02 -2.80944586e-01
6.81822836e-01 4.63556647e-01 -3.90825629e-01 -2.91109532e-01
-4.59926277e-01 -1.02907479e-01 -9.34488028e-02 5.51751912e-01
3.96546513e-01 4.78258163e-01 -2.48174682e-01 -2.94033512e-02
-2.26112753e-02 -4.13408697e-01 8.76650810e-01 7.94486225e-01
-5.31296194e-01 -1.31043291e+00 -2.88753271e-01 6.02638721e-01
-3.07708949e-01 -2.37659946e-01 2.12292582e-01 3.37891817e-01
-3.03706378e-01 6.60704434e-01 -1.03185944e-01 -2.66514510e-01
-3.54656996e-03 7.62150288e-02 1.10687506e+00 -4.35016245e-01
-1.62615970e-01 -1.85297817e-01 1.14497907e-01 -6.29509091e-01
-4.28080797e-01 -6.76016152e-01 -7.81406760e-01 -3.46813172e-01
-7.84922063e-01 4.28444475e-01 1.03094828e+00 1.04989803e+00
4.79302377e-01 4.38415617e-01 6.08569443e-01 -1.19660175e+00
-1.27226532e+00 -1.07557845e+00 -1.16718125e+00 -2.37162396e-01
3.02131802e-01 -8.37969899e-01 -4.55104381e-01 -7.92547345e-01]
|
[5.6091508865356445, 4.914169788360596]
|
5bd2403a-06f8-48cf-95b2-c4ef3fc41bc3
|
learning-towards-abstractive-timeline
| null | null |
https://www.ijcai.org/proceedings/2019/686
|
https://www.ijcai.org/proceedings/2019/0686.pdf
|
Learning towards Abstractive Timeline Summarization
|
Timeline summarization targets at concisely summarizing the evolution trajectory along the timeline and existing timeline summarization approaches are all based on extractive methods.In this paper, we propose the task of abstractive timeline summarization, which tends to concisely paraphrase the information in the time-stamped events.Unlike traditional document summarization, timeline summarization needs to model the time series information of the input events and summarize important events in chronological order.To tackle this challenge, we propose a memory-based timeline summarization model (MTS).Concretely, we propose a time-event memory to establish a timeline, and use the time position of events on this timeline to guide generation process.Besides, in each decoding step, we incorporate event-level information into word-level attention to avoid confusion between events.Extensive experiments are conducted on a large-scale real-world dataset, and the results show that MTS achieves the state-of-the-art performance in terms of both automatic and human evaluations.
|
['Meng-Hsuan Yu', 'Zhangming Chan', 'Shen Gao', 'Xiuying Chen', 'Rui Yan', 'Dongyan Zhao']
|
2019-08-11
| null | null | null |
ijcai-2019-2019-8
|
['timeline-summarization']
|
['natural-language-processing']
|
[ 4.54616696e-01 -1.10497892e-01 -3.51498663e-01 -1.90151662e-01
-1.05200756e+00 -3.87702286e-01 8.52986157e-01 9.93989050e-01
-3.21861923e-01 7.98810840e-01 1.32123232e+00 -6.39997050e-02
-7.78426081e-02 -6.18694246e-01 -5.67547083e-01 -1.56975329e-01
-1.17527865e-01 2.36353278e-01 1.01451106e-01 -5.90731064e-03
7.91685164e-01 1.01692580e-01 -1.12018561e+00 4.89432663e-01
1.30402744e+00 5.02948880e-01 1.05606295e-01 7.88439453e-01
-2.95924515e-01 9.52112019e-01 -1.36533260e+00 -2.22398087e-01
-3.74289960e-01 -9.16155994e-01 -7.05973864e-01 2.84646731e-02
2.93480307e-01 -3.31160396e-01 -8.06761086e-01 7.49580801e-01
5.38051069e-01 4.79331255e-01 5.48049092e-01 -1.04953134e+00
-4.04775470e-01 1.00951314e+00 -8.04886818e-01 7.03435123e-01
6.63447559e-01 -9.10003558e-02 1.08276176e+00 -4.73427743e-01
6.22901857e-01 1.10888875e+00 4.91279960e-01 1.55961707e-01
-6.90494895e-01 -3.24925482e-01 5.77725232e-01 3.73922199e-01
-8.68131697e-01 -2.95172542e-01 8.86321664e-01 -7.21909255e-02
1.41777015e+00 5.54255486e-01 9.00665224e-01 7.66067505e-01
5.20888269e-01 1.18550789e+00 1.60625741e-01 -2.67593235e-01
1.83770195e-01 -8.23652148e-01 5.24866760e-01 1.18263856e-01
1.97049886e-01 -5.82490802e-01 -8.18967581e-01 -2.08646566e-01
2.16605335e-01 1.72721103e-01 -2.57015318e-01 4.60435331e-01
-1.57840848e+00 6.10170186e-01 2.07008943e-01 3.64818066e-01
-6.91789150e-01 1.35673076e-01 1.13877475e+00 1.39826015e-01
9.27031398e-01 5.72772145e-01 -7.19220787e-02 -4.60420281e-01
-1.54376423e+00 4.28134292e-01 7.86908448e-01 8.38252425e-01
1.15523182e-01 6.04901910e-02 -9.48702514e-01 6.94502771e-01
-3.10634285e-01 2.49789357e-01 8.10853183e-01 -6.13910675e-01
9.25122380e-01 7.51338542e-01 2.31186777e-01 -1.11637819e+00
-3.58943641e-01 -4.89570826e-01 -9.38029230e-01 -8.54465544e-01
-2.91280210e-01 -2.90976137e-01 -8.02282035e-01 1.47166109e+00
6.64382875e-02 5.73803306e-01 4.35786694e-02 4.78085399e-01
1.03152025e+00 1.49910033e+00 9.18611288e-02 -1.12898588e+00
1.22054648e+00 -1.20640850e+00 -1.09311688e+00 -1.84009209e-01
4.43624735e-01 -5.48818648e-01 8.02556336e-01 8.44352618e-02
-1.40926778e+00 -4.58498031e-01 -1.27390957e+00 -2.08398312e-01
-9.13023427e-02 1.24107793e-01 2.74208724e-01 -7.21153915e-02
-6.31267548e-01 7.48958945e-01 -1.12047648e+00 -5.53975224e-01
8.93419757e-02 -1.98511377e-01 -9.67824161e-02 3.52985173e-01
-1.25412500e+00 8.04633975e-01 9.78519499e-01 -8.32632333e-02
-5.73905110e-01 -8.83860767e-01 -9.48628724e-01 2.73050129e-01
3.20583820e-01 -8.01161706e-01 1.67502666e+00 -1.85852230e-01
-1.16228485e+00 4.26314652e-01 -7.50637412e-01 -8.11355352e-01
4.06053960e-01 -5.80928385e-01 -5.74625671e-01 3.94158989e-01
3.23176533e-01 2.25370511e-01 4.76129025e-01 -6.97557151e-01
-9.60761905e-01 -2.25944780e-02 -1.59488186e-01 3.71211141e-01
-4.24838722e-01 2.67311215e-01 -6.86371565e-01 -1.05003059e+00
-7.06006140e-02 -2.24084884e-01 -3.31127226e-01 -9.01445210e-01
-8.75799119e-01 -5.19390762e-01 6.90827787e-01 -1.02277565e+00
2.43646479e+00 -1.89433920e+00 3.74606907e-01 -4.10180807e-01
1.60763547e-01 2.47023165e-01 1.89737976e-02 1.10447967e+00
3.01470724e-03 1.13032050e-01 -5.04333019e-01 -5.27403951e-01
1.04343891e-02 -1.05254389e-01 -8.76702607e-01 2.24979222e-01
2.31951118e-01 1.04423535e+00 -1.26719749e+00 -7.18603253e-01
1.66791230e-01 5.00094071e-02 -2.93212891e-01 2.99571931e-01
-5.13898075e-01 3.01076472e-01 -5.36249280e-01 2.18341246e-01
2.44777739e-01 -6.03915974e-02 -2.32645929e-01 -2.86499590e-01
-4.31456596e-01 7.84601450e-01 -5.94231129e-01 1.92736459e+00
-3.07041556e-01 6.83732569e-01 -8.09486747e-01 -8.08126926e-01
7.10304558e-01 3.14589053e-01 6.49966478e-01 -7.18728960e-01
-3.74547578e-02 8.07594322e-03 -4.48747814e-01 -4.14096296e-01
1.26971352e+00 2.74605632e-01 -6.12159610e-01 7.88082898e-01
-8.34300593e-02 -1.78765982e-01 9.22544241e-01 5.72174788e-01
1.07857490e+00 -1.17256768e-01 8.03051710e-01 2.48914465e-01
3.78717899e-01 2.78360188e-01 4.96909738e-01 6.97803855e-01
1.53171539e-01 7.92747438e-01 7.73074329e-01 -2.95056641e-01
-1.16784310e+00 -8.46513271e-01 3.35760504e-01 9.21297312e-01
6.08599409e-02 -1.19986725e+00 -6.63065791e-01 -5.42946517e-01
-3.63933146e-01 1.35302114e+00 -5.07446587e-01 -3.51785034e-01
-1.03867018e+00 -8.96569610e-01 4.98588055e-01 5.70629179e-01
4.50794369e-01 -1.20660949e+00 -8.65603745e-01 7.84927785e-01
-6.34903729e-01 -7.11054146e-01 -1.02744174e+00 -2.21928656e-01
-8.26365590e-01 -6.77520335e-01 -8.65898788e-01 -6.38022363e-01
4.61662918e-01 2.33352989e-01 1.09270620e+00 -2.34553501e-01
3.53394374e-02 1.33541912e-01 -5.04457355e-01 -7.57749856e-01
-2.01257825e-01 4.57493156e-01 -3.61711144e-01 -1.71348065e-01
1.90162838e-01 -6.82669938e-01 -5.78454554e-01 -2.57079095e-01
-1.07359076e+00 4.68940020e-01 2.53097504e-01 5.63223600e-01
5.93416810e-01 1.12784199e-01 1.09576368e+00 -7.47236013e-01
1.21556437e+00 -7.15353847e-01 -4.53201681e-02 4.82547194e-01
-1.79764435e-01 -1.57654569e-01 8.05380046e-01 -4.22315419e-01
-1.01647973e+00 -4.90454614e-01 -2.24644840e-02 5.52322939e-02
2.65739024e-01 1.10628676e+00 5.18420599e-02 1.11910272e+00
3.05454224e-01 7.89788961e-01 -5.89319408e-01 -4.03274953e-01
5.61608970e-01 7.10644901e-01 9.61704016e-01 -3.75909775e-01
3.51755261e-01 3.93306077e-01 -4.73850250e-01 -7.04698801e-01
-1.11772323e+00 -6.10624254e-01 -4.23338085e-01 -3.50564063e-01
5.55915773e-01 -8.17379355e-01 -1.60747215e-01 3.22821200e-01
-1.53866005e+00 2.94690710e-02 -4.86448854e-01 3.37437242e-01
-5.70682764e-01 6.34297132e-01 -5.85172772e-01 -5.69174469e-01
-9.00645137e-01 -3.83866876e-01 1.26020873e+00 6.48366511e-01
-8.09576929e-01 -8.72785985e-01 2.71256059e-01 -2.89946020e-01
2.16623858e-01 6.04086041e-01 8.80022585e-01 -9.33800578e-01
-1.44987196e-01 -5.02920270e-01 -5.98339178e-02 -2.49214679e-01
3.42078865e-01 1.91570073e-01 -3.31348360e-01 -1.15980528e-01
1.00282513e-01 7.97803551e-02 1.26298714e+00 5.34878790e-01
1.16313994e+00 -6.91824615e-01 -4.28777188e-01 2.89660931e-01
9.06362653e-01 4.23826277e-01 5.88164449e-01 2.84098566e-01
7.18924880e-01 5.82651973e-01 8.77047420e-01 8.42403650e-01
6.64563954e-01 3.25880796e-01 -3.33141014e-02 1.25134155e-01
-1.36478413e-02 -7.03017890e-01 3.88890564e-01 1.46141279e+00
2.71390676e-01 -7.53401041e-01 -6.76076055e-01 9.94591475e-01
-2.34534979e+00 -1.43721473e+00 -1.43919185e-01 2.06326318e+00
9.73423660e-01 2.51480907e-01 2.02844501e-01 1.76804915e-01
8.30312073e-01 9.20393109e-01 -7.13727117e-01 -5.12362003e-01
-1.09175630e-01 -2.86192864e-01 4.54066508e-02 3.57720494e-01
-1.02754211e+00 8.25803280e-01 6.13940334e+00 9.31201577e-01
-1.10223711e+00 -7.96501264e-02 5.67808628e-01 -5.62151968e-01
-4.67012644e-01 -1.59563318e-01 -5.85842550e-01 7.50598252e-01
1.13333893e+00 -1.34837425e+00 -2.83085499e-02 3.07847142e-01
7.14299381e-01 -1.79438770e-01 -1.19147646e+00 8.98571670e-01
3.53838354e-01 -1.51494789e+00 5.42313933e-01 -5.77949584e-01
8.60182703e-01 -2.90571541e-01 -4.31945503e-01 4.13646221e-01
3.96946408e-02 -7.04018354e-01 8.00816536e-01 5.87280869e-01
6.22488499e-01 -9.80391622e-01 4.82708037e-01 4.40599114e-01
-1.47079480e+00 2.07661197e-01 -1.44890904e-01 -5.34978807e-02
9.50998068e-01 7.30787337e-01 -6.12132013e-01 1.17082012e+00
1.41646549e-01 1.38204217e+00 -4.27610606e-01 1.43345726e+00
-3.01296115e-01 9.14474785e-01 -2.71482289e-01 -1.16726868e-01
2.74081171e-01 -3.46498489e-02 9.61909473e-01 1.62788308e+00
5.47422707e-01 2.33488411e-01 1.53589696e-01 2.56995857e-01
-1.39156222e-01 7.48127550e-02 -3.07239234e-01 -4.41315591e-01
5.83618760e-01 8.28554809e-01 -8.09988856e-01 -7.30404556e-01
-3.46832536e-02 1.10513186e+00 1.29268289e-01 2.14338630e-01
-1.04826367e+00 -8.35093737e-01 1.52555453e-02 -1.94621935e-01
2.09198639e-01 -2.71301419e-01 -2.12640122e-01 -1.28134501e+00
3.07618618e-01 -5.43615639e-01 7.95286298e-01 -9.51017439e-01
-1.10869539e+00 5.68967402e-01 3.22656155e-01 -1.36947227e+00
-4.01462495e-01 5.13483882e-01 -1.38575828e+00 5.65140367e-01
-1.27532518e+00 -8.57729495e-01 -1.27943367e-01 4.83197086e-02
1.24055111e+00 1.55520022e-01 3.59270632e-01 1.57391131e-01
-7.40344763e-01 1.90688878e-01 1.54708222e-01 -1.34530678e-01
6.74787581e-01 -1.20895326e+00 1.07517779e+00 1.20022571e+00
-1.37522668e-02 7.42097318e-01 9.82791424e-01 -1.09681606e+00
-1.20838678e+00 -1.44238544e+00 1.61302865e+00 -1.72107518e-01
7.06775427e-01 7.84239396e-02 -9.57445621e-01 6.80443645e-01
7.06088543e-01 -8.74035954e-01 4.99766201e-01 -1.65755019e-01
-2.93158293e-02 4.20376845e-02 -5.36862016e-01 9.96583760e-01
1.16283286e+00 -2.81781733e-01 -1.36665606e+00 5.32433510e-01
1.16600478e+00 -6.08829141e-01 -5.22088766e-01 3.06918919e-01
2.14373425e-01 -4.17785078e-01 6.54796362e-01 -6.33176148e-01
8.58974099e-01 -4.39562470e-01 3.73377472e-01 -1.70988524e+00
-3.18428934e-01 -1.11830270e+00 -5.19375563e-01 1.57317865e+00
1.54891282e-01 -3.25324863e-01 3.08570445e-01 -1.25552816e-02
-6.08353674e-01 -6.37240589e-01 -7.82613337e-01 -5.25692463e-01
-3.53150338e-01 -2.33592391e-01 7.32530355e-01 6.31137192e-01
5.41818023e-01 7.24348247e-01 -5.55410564e-01 -1.12916872e-01
3.47129136e-01 4.91551340e-01 5.97928524e-01 -8.69926035e-01
1.69987500e-01 -6.53751373e-01 6.04511648e-02 -1.21987748e+00
1.14820912e-01 -6.42825305e-01 9.41294581e-02 -2.29162645e+00
4.10512954e-01 4.41752851e-01 -2.10528210e-01 1.60177514e-01
-7.84846783e-01 -3.80148411e-01 1.67995915e-01 3.01847875e-01
-1.25377607e+00 9.11447287e-01 1.03594983e+00 -3.90109152e-01
-4.52351898e-01 -8.43294784e-02 -8.29692483e-01 4.84840870e-01
7.86905348e-01 -5.39683700e-01 -6.49782002e-01 -4.92950439e-01
1.81980774e-01 5.07843316e-01 -1.56324670e-01 -1.04371381e+00
7.51493633e-01 -2.42553681e-01 -2.39834692e-02 -1.49013793e+00
7.73774609e-02 -1.74183145e-01 1.63501307e-01 4.17416036e-01
-7.38737166e-01 4.63709056e-01 3.15367073e-01 7.16113985e-01
-5.93888938e-01 -1.46317884e-01 1.87130436e-01 6.41796067e-02
-5.76507688e-01 2.30609775e-01 -3.95857543e-01 2.50811189e-01
9.17890966e-01 -1.79502562e-01 -6.08452141e-01 -5.49773574e-01
-2.47710243e-01 7.89411247e-01 1.46297604e-01 4.49246734e-01
5.20731807e-01 -1.26800680e+00 -1.19422293e+00 -3.84967953e-01
1.73018396e-01 1.55972049e-01 4.72455800e-01 6.29699588e-01
-5.53471625e-01 6.35036945e-01 -3.62337008e-02 -3.39124292e-01
-1.09883451e+00 5.19689977e-01 -1.56951472e-01 -5.63032806e-01
-8.98657501e-01 4.92978364e-01 3.44399139e-02 3.27924222e-01
3.16689909e-01 -6.53578579e-01 -6.57528698e-01 5.89992225e-01
1.20266283e+00 4.81919497e-01 7.04184324e-02 -3.49818379e-01
-4.28363644e-02 3.67543340e-01 -5.06038249e-01 -1.97717026e-01
1.47187436e+00 -1.80366427e-01 -2.94788748e-01 8.33608747e-01
9.64420974e-01 1.14186488e-01 -1.00317585e+00 -2.08571643e-01
3.75215352e-01 -9.71106440e-02 -2.64655948e-01 -6.14682198e-01
-4.79475170e-01 5.49635231e-01 -5.32415688e-01 4.64266568e-01
1.36075318e+00 -1.47638306e-01 1.37700236e+00 2.29854330e-01
-5.64302951e-02 -9.73146975e-01 -6.47977516e-02 8.20233047e-01
1.21546674e+00 -6.28951728e-01 3.15668464e-01 -2.04682186e-01
-6.67565405e-01 1.04828560e+00 3.88467818e-01 -1.54483378e-01
1.43823819e-02 -1.53350472e-01 -4.31650490e-01 -5.37001789e-02
-1.11192954e+00 -1.05622527e-03 2.88524717e-01 7.65893087e-02
4.43578392e-01 -5.05624302e-02 -9.94473279e-01 8.40663970e-01
-4.37824130e-01 -7.74991214e-02 6.64537668e-01 1.04483593e+00
-5.73337197e-01 -7.09665120e-01 -2.49834180e-01 7.06229329e-01
-4.76838559e-01 -7.88139105e-02 -5.29559851e-01 3.21859241e-01
-5.46369076e-01 1.13543189e+00 3.04746091e-01 -1.03524819e-01
7.01843560e-01 -2.15421785e-02 1.45442218e-01 -7.63730049e-01
-9.02005732e-01 1.39113739e-01 3.28724384e-01 -2.70707577e-01
-2.48639569e-01 -7.72697687e-01 -1.41287839e+00 -3.55763137e-01
5.74987717e-02 5.77988803e-01 3.74685675e-01 8.61487627e-01
5.53364933e-01 1.09327769e+00 6.63831592e-01 -9.23222423e-01
-2.70970970e-01 -1.22268701e+00 -1.43466309e-01 3.24720263e-01
5.00569403e-01 -5.79544157e-02 -1.52180731e-01 1.72558725e-01]
|
[12.56629467010498, 9.490503311157227]
|
f23c8539-039b-480d-ac9a-571d35a340a2
|
text-adaptive-multiple-visual-prototype
|
2209.13307
| null |
https://arxiv.org/abs/2209.13307v1
|
https://arxiv.org/pdf/2209.13307v1.pdf
|
Text-Adaptive Multiple Visual Prototype Matching for Video-Text Retrieval
|
Cross-modal retrieval between videos and texts has gained increasing research interest due to the rapid emergence of videos on the web. Generally, a video contains rich instance and event information and the query text only describes a part of the information. Thus, a video can correspond to multiple different text descriptions and queries. We call this phenomenon the ``Video-Text Correspondence Ambiguity'' problem. Current techniques mostly concentrate on mining local or multi-level alignment between contents of a video and text (\textit{e.g.}, object to entity and action to verb). It is difficult for these methods to alleviate the video-text correspondence ambiguity by describing a video using only one single feature, which is required to be matched with multiple different text features at the same time. To address this problem, we propose a Text-Adaptive Multiple Visual Prototype Matching model, which automatically captures multiple prototypes to describe a video by adaptive aggregation of video token features. Given a query text, the similarity is determined by the most similar prototype to find correspondence in the video, which is termed text-adaptive matching. To learn diverse prototypes for representing the rich information in videos, we propose a variance loss to encourage different prototypes to attend to different contents of the video. Our method outperforms state-of-the-art methods on four public video retrieval datasets.
|
['Chunhua Shen', 'Wei-Shi Zheng', 'Wenhang Ge', 'Jun Zhang', 'Junwei Liang', 'AnCong Wu', 'Chengzhi Lin']
|
2022-09-27
| null | null | null | null |
['video-text-retrieval']
|
['computer-vision']
|
[ 1.63772076e-01 -7.25697875e-01 -5.00941634e-01 -4.12596703e-01
-8.05456161e-01 -6.67544305e-01 5.87358415e-01 4.79510397e-01
-3.15655857e-01 1.89762875e-01 4.32693064e-01 5.51309288e-01
-1.85110912e-01 -3.20441872e-01 -6.54647171e-01 -4.92218345e-01
1.14299871e-01 3.94223332e-01 5.20899653e-01 1.16358921e-01
3.45708311e-01 6.03257306e-02 -1.98039174e+00 9.23071802e-01
3.67304057e-01 1.22680414e+00 6.27840221e-01 3.25580925e-01
-4.96389359e-01 6.89039886e-01 -4.97560859e-01 -2.57844836e-01
2.18736649e-01 -5.63155472e-01 -6.15085185e-01 4.44430977e-01
9.06551838e-01 -4.00232643e-01 -5.58868408e-01 1.05785561e+00
2.83626676e-01 3.11898530e-01 5.57800591e-01 -1.67503369e+00
-4.34955508e-01 5.10067701e-01 -6.83074355e-01 3.12858224e-01
9.20604289e-01 -3.43850315e-01 1.15716565e+00 -9.66257572e-01
9.37889755e-01 1.27355957e+00 3.30865711e-01 3.77614200e-01
-8.85424376e-01 -5.84227145e-01 2.09004045e-01 7.61374176e-01
-1.83835471e+00 -4.88689363e-01 6.74425602e-01 -4.64065403e-01
6.63689315e-01 4.10095453e-01 7.36433923e-01 9.70284641e-01
5.00970185e-02 1.07421863e+00 4.66758043e-01 -2.28763014e-01
-1.10288113e-01 1.97520524e-01 -4.20386828e-02 6.15605474e-01
6.08152933e-02 -5.44510543e-01 -8.26260328e-01 -2.80632943e-01
4.11375940e-01 4.32477117e-01 -4.57620949e-01 -3.96391332e-01
-1.29138398e+00 6.97176993e-01 -5.19101806e-02 4.69162405e-01
-3.94986123e-01 4.84604202e-02 8.02859545e-01 3.89762938e-01
4.08525877e-02 1.50417477e-01 -2.41786242e-01 -1.62603542e-01
-1.06903136e+00 4.37517524e-01 5.83955646e-01 1.37786961e+00
9.92699265e-01 -5.69330692e-01 -3.07820290e-01 8.16490412e-01
3.13240200e-01 3.82385015e-01 8.27419460e-01 -9.30439770e-01
6.58982277e-01 7.32574284e-01 -3.84837352e-02 -1.42658377e+00
1.23276664e-02 2.00034410e-01 -5.64093292e-01 -6.56806290e-01
1.10486589e-01 4.67869937e-01 -5.55000544e-01 1.45714045e+00
3.68399650e-01 3.93698692e-01 -1.39438063e-01 9.41399872e-01
1.06126499e+00 9.81492698e-01 1.46031240e-02 -6.13420725e-01
1.40926218e+00 -8.10360253e-01 -7.90819168e-01 -7.28077367e-02
5.34208059e-01 -1.07585049e+00 8.38046074e-01 5.65769486e-02
-9.63819504e-01 -5.88183820e-01 -7.09550858e-01 4.45880480e-02
-3.10824722e-01 -3.68549600e-02 -1.38282239e-01 4.22220230e-02
-6.00660741e-01 4.16087776e-01 -3.50104809e-01 -7.22745419e-01
1.05610326e-01 6.26680851e-02 -6.10272706e-01 -5.31095445e-01
-1.17108703e+00 3.86329353e-01 7.08353162e-01 -4.33098286e-01
-5.44088364e-01 -4.63047117e-01 -8.79444957e-01 1.65715422e-02
8.17881465e-01 -5.67453265e-01 9.36613321e-01 -1.35222316e+00
-8.66156757e-01 7.75868177e-01 -5.47233939e-01 -7.41167516e-02
3.21454912e-01 -1.45606950e-01 -6.76563203e-01 6.89392924e-01
4.56261277e-01 6.61081553e-01 1.38668501e+00 -1.09387457e+00
-1.01334453e+00 -1.73095331e-01 -2.34981086e-02 3.53965163e-01
-6.28886044e-01 4.87859249e-01 -1.20620096e+00 -8.66536975e-01
2.64642745e-01 -7.62915313e-01 4.53906327e-01 1.49643108e-01
-9.37005877e-02 -5.56477964e-01 1.09803820e+00 -5.94682693e-01
1.66593730e+00 -2.27322340e+00 3.81675899e-01 2.16051757e-01
1.84859350e-01 -7.61736631e-02 -2.47641265e-01 6.63384497e-01
9.22730193e-02 1.40344024e-01 1.55868754e-01 -9.69988331e-02
-1.00425862e-01 2.30559245e-01 -2.36608654e-01 5.06099939e-01
-1.39494002e-01 5.23810089e-01 -9.74999964e-01 -1.20984662e+00
3.11641116e-02 3.73107821e-01 -5.01528144e-01 2.48236239e-01
-2.45884851e-01 3.60012725e-02 -6.24850929e-01 7.63809144e-01
3.16227585e-01 -3.20697695e-01 1.93901300e-01 -7.54846036e-01
-1.17896989e-01 -2.75589496e-01 -1.32885313e+00 1.84302223e+00
1.16134239e-02 7.65470982e-01 -2.12986141e-01 -1.08736622e+00
4.83773291e-01 6.12436116e-01 8.91360819e-01 -6.37592852e-01
-8.46290737e-02 1.85302228e-01 -3.71276945e-01 -1.00860703e+00
5.63431382e-01 1.84507087e-01 -1.37341887e-01 3.94876689e-01
-7.26460814e-02 2.63650596e-01 5.06818831e-01 3.66264850e-01
9.09131050e-01 1.19485058e-01 3.84844363e-01 1.26552954e-01
6.54466331e-01 1.08996056e-01 6.93467557e-01 8.20086122e-01
-1.83201090e-01 6.11838758e-01 2.47297689e-01 -3.57793748e-01
-1.10219359e+00 -6.16283834e-01 2.12702565e-02 1.26929057e+00
5.31620204e-01 -9.97421741e-01 -4.50186998e-01 -5.63264310e-01
1.61365986e-01 -1.12716593e-02 -4.72210258e-01 1.17054125e-02
-6.76697254e-01 3.85238118e-02 3.12808484e-01 1.81753963e-01
4.33527827e-01 -7.99339414e-01 -4.67505872e-01 2.48280004e-01
-7.25583017e-01 -1.26958072e+00 -1.07917976e+00 -5.09436071e-01
-4.67170089e-01 -1.19485879e+00 -7.43323386e-01 -9.09520566e-01
5.11651337e-01 7.52434671e-01 1.00494313e+00 2.68924415e-01
-4.31075752e-01 7.14143515e-01 -9.11335647e-01 2.05816552e-01
-3.31862122e-01 -2.39767954e-01 2.00319931e-01 3.07412773e-01
5.69145143e-01 -1.43675972e-02 -5.66308141e-01 6.39532864e-01
-1.31985843e+00 1.31178796e-02 3.73433620e-01 7.42972553e-01
8.63009751e-01 1.80558160e-01 1.39193326e-01 -2.30342224e-01
3.34481150e-01 -8.01151931e-01 -1.20503955e-01 7.14794397e-01
-1.90596372e-01 -9.07327309e-02 6.67010009e-01 -8.56637359e-01
-4.12403494e-01 6.83541521e-02 5.93815804e-01 -1.15632451e+00
-1.56072229e-01 7.58218944e-01 -1.36887252e-01 2.00515360e-01
1.72087416e-01 5.32833934e-01 -2.83133537e-01 -2.86761403e-01
1.31714329e-01 8.27678859e-01 3.20803821e-01 -5.84307849e-01
6.83646500e-01 3.17808211e-01 -9.15226266e-02 -1.02283835e+00
-6.80299222e-01 -1.14215171e+00 -5.45027792e-01 -5.69214225e-01
8.08787465e-01 -9.42143083e-01 -4.12310421e-01 1.39497846e-01
-1.12516236e+00 3.79891098e-01 3.80211659e-02 5.91290593e-01
-5.10357499e-01 8.60938668e-01 -2.46630877e-01 -2.97697484e-01
-1.32175043e-01 -1.13102698e+00 1.14925075e+00 7.52898157e-02
-2.69318312e-01 -5.80728471e-01 -6.31388053e-02 2.99232483e-01
6.56982288e-02 -2.78585069e-02 8.41397524e-01 -8.25163901e-01
-5.13786912e-01 -4.23323721e-01 -2.59966224e-01 -3.93480100e-02
3.31689924e-01 2.63086408e-01 -4.04953688e-01 -5.85792542e-01
-1.78242832e-01 -3.55612338e-01 5.98040879e-01 2.01167628e-01
1.08969712e+00 -6.05757415e-01 -5.97330034e-01 3.64246428e-01
1.40047419e+00 2.84417719e-01 3.20490181e-01 4.34311152e-01
8.29060435e-01 6.19186461e-01 9.58247423e-01 7.53593326e-01
2.93517113e-01 1.14735532e+00 1.41698897e-01 5.30266464e-01
1.04282148e-01 -3.20176750e-01 3.89252007e-01 9.04284716e-01
2.78059155e-01 -4.33479518e-01 -6.45921052e-01 5.55666327e-01
-2.11891341e+00 -1.57424903e+00 2.26147577e-01 2.26003861e+00
5.91807783e-01 -2.98352033e-01 2.29563549e-01 -5.49306422e-02
1.19550514e+00 2.08028316e-01 -4.64519560e-01 1.87965617e-01
-1.50168628e-01 -5.02839506e-01 6.02636486e-02 -2.68949959e-02
-1.08947957e+00 7.17442274e-01 5.00868082e+00 1.10165679e+00
-8.99518371e-01 1.03228623e-02 1.26042413e-02 -3.16168427e-01
-2.58980751e-01 1.93092506e-02 -8.18589866e-01 7.41528451e-01
5.45423567e-01 -5.01926363e-01 3.74179274e-01 6.12110972e-01
3.36755693e-01 3.99489701e-03 -1.29795134e+00 1.47481561e+00
7.15047061e-01 -1.30954111e+00 7.65891135e-01 -2.43618801e-01
5.97841084e-01 -3.06904912e-01 -1.63876206e-01 1.37041032e-01
-5.49369454e-01 -5.05443990e-01 9.60157692e-01 4.24137563e-01
7.95018137e-01 -5.87764025e-01 4.10939842e-01 2.01263368e-01
-1.92405951e+00 -1.28650263e-01 -4.36857879e-01 6.08338833e-01
8.41598213e-02 8.71471018e-02 -5.15127182e-01 6.34242833e-01
9.74125862e-01 1.21633542e+00 -5.03068864e-01 1.30577409e+00
4.98341233e-01 2.29234528e-02 -2.95942038e-01 2.54746042e-02
3.80985677e-01 -9.39970165e-02 6.89188063e-01 1.17009830e+00
5.31950772e-01 6.43209964e-02 7.38955081e-01 4.28242445e-01
-2.30876654e-01 6.23542905e-01 -6.39677346e-01 -2.11425394e-01
8.14057469e-01 9.73602295e-01 -6.71302497e-01 -5.97201765e-01
-8.61967504e-01 1.16581297e+00 6.73914328e-02 1.96727395e-01
-8.06077719e-01 -3.27067405e-01 6.30094945e-01 1.61485389e-01
4.01910573e-01 8.34205970e-02 7.62955308e-01 -1.33076775e+00
3.24217051e-01 -9.49985385e-01 8.12820077e-01 -9.57597554e-01
-1.36129677e+00 4.78343964e-01 3.72536957e-01 -1.94960773e+00
-3.34243029e-01 -1.21395350e-01 -1.92641050e-01 4.16314960e-01
-1.24123847e+00 -1.01695359e+00 -4.84223545e-01 1.12563455e+00
1.03807986e+00 -2.90584266e-01 4.63101208e-01 5.33890367e-01
-3.63042921e-01 5.69142938e-01 3.42959732e-01 2.02933475e-01
1.16013050e+00 -5.36270618e-01 -1.46356463e-01 6.23694003e-01
2.75861770e-01 4.84764427e-01 7.43106008e-01 -8.42543721e-01
-1.79408717e+00 -1.11190116e+00 9.80210125e-01 -9.68814790e-02
8.99911940e-01 1.00719154e-01 -9.74551737e-01 7.03058839e-01
4.43735905e-02 1.13145091e-01 7.88859904e-01 -5.84000409e-01
-5.12179673e-01 -1.68698803e-01 -9.51215148e-01 5.90057194e-01
9.67148304e-01 -7.68851221e-01 -7.15037763e-01 4.55891967e-01
6.74993753e-01 -1.83020398e-01 -9.40244079e-01 1.58760786e-01
7.67093778e-01 -7.17549920e-01 8.39406252e-01 -6.49353385e-01
3.33493918e-01 -4.87844318e-01 -5.95095873e-01 -9.08442914e-01
-2.36650348e-01 -5.42232811e-01 -2.37863794e-01 1.29015183e+00
-6.78140670e-02 1.08758425e-02 5.70937634e-01 4.25151438e-01
-3.80914621e-02 -3.66167039e-01 -1.14229465e+00 -8.72166753e-01
-4.71271545e-01 -2.09114194e-01 5.05847275e-01 1.22995210e+00
3.07782739e-01 2.23248079e-03 -6.13941610e-01 6.06622845e-02
5.15286446e-01 3.88247818e-01 5.71657419e-01 -1.21480322e+00
-4.29007150e-02 -4.82761502e-01 -7.94874370e-01 -1.12185001e+00
3.87255043e-01 -9.41735744e-01 7.23044798e-02 -1.25153756e+00
7.90069342e-01 -1.05242386e-01 -5.68679050e-02 2.97182649e-01
-8.39390084e-02 1.41171470e-01 3.86490464e-01 8.27275038e-01
-1.16292584e+00 3.39919984e-01 9.66404378e-01 -4.15649086e-01
-7.75503069e-02 -3.07358652e-01 -1.91192880e-01 5.01789927e-01
4.23629016e-01 -6.37539208e-01 -3.05844724e-01 -5.18089652e-01
3.03982764e-01 3.66055757e-01 2.32089117e-01 -9.24793541e-01
5.55712342e-01 -3.75356644e-01 2.55585015e-01 -6.40391588e-01
3.24145854e-01 -1.27828383e+00 4.09854561e-01 2.16668218e-01
-4.91957039e-01 4.19348389e-01 -1.24480031e-01 8.63824308e-01
-5.95841765e-01 -3.54481548e-01 5.04946709e-01 -2.96393484e-01
-1.18216181e+00 6.45221055e-01 -4.33101654e-01 1.00339264e-01
1.15305543e+00 -5.40508807e-01 -3.77785176e-01 -3.72834563e-01
-4.29432034e-01 4.53586757e-01 7.30886698e-01 8.18270743e-01
9.70805943e-01 -1.60280800e+00 -7.13203490e-01 -5.84521927e-02
7.58712232e-01 -3.82281601e-01 3.50121856e-01 6.33632720e-01
-2.58787572e-01 3.13483804e-01 -1.11303329e-01 -8.53666961e-01
-1.83375418e+00 7.75208414e-01 1.01198740e-01 2.45211869e-01
-7.32777596e-01 4.72159386e-01 1.66178718e-02 4.27126884e-01
3.98592591e-01 1.89398741e-03 -4.60482299e-01 4.72035140e-01
7.99951494e-01 2.34129101e-01 -2.53630400e-01 -1.31400359e+00
-4.41211939e-01 1.05731666e+00 -2.83813477e-01 1.84837192e-01
8.52439642e-01 -5.14179647e-01 -1.91448927e-01 6.02834463e-01
1.61960995e+00 -2.64225960e-01 -8.62083435e-01 -6.67392254e-01
-1.90638360e-02 -9.32906449e-01 -1.45753562e-01 -1.59145057e-01
-1.05621159e+00 3.29093516e-01 5.75534105e-01 2.62814224e-01
1.01909268e+00 3.34767759e-01 9.12710965e-01 5.97599447e-01
2.40057081e-01 -1.25931811e+00 3.63572180e-01 2.40253329e-01
7.17974663e-01 -1.33858299e+00 1.60131380e-01 -1.70377880e-01
-7.72459090e-01 1.20587325e+00 5.67193329e-01 3.17286134e-01
4.66762453e-01 -2.45178685e-01 -2.37870619e-01 -2.54849344e-01
-7.68701911e-01 -1.00370035e-01 6.06029212e-01 1.94125578e-01
1.87129661e-01 -2.86510348e-01 -2.91089237e-01 7.01989755e-02
4.92265671e-01 -1.18807033e-01 2.76741862e-01 1.17039788e+00
-6.78062379e-01 -1.03802884e+00 -3.77518177e-01 6.56804800e-01
-4.67750937e-01 -2.32046619e-02 -2.71018982e-01 6.96305454e-01
-2.74384376e-02 9.61664200e-01 3.88226628e-01 -4.54690278e-01
2.05912828e-01 1.20284110e-01 3.64639044e-01 -5.49946845e-01
-3.99346441e-01 3.72796148e-01 -3.16623747e-01 -7.20898092e-01
-9.02816713e-01 -1.05349982e+00 -1.15646434e+00 -3.41703385e-01
-2.38801450e-01 3.27794790e-01 3.20399165e-01 1.02561069e+00
4.03341323e-01 -4.86316867e-02 7.97851503e-01 -7.86624432e-01
-1.68355331e-01 -6.30652845e-01 -5.96989453e-01 9.61950362e-01
2.01933369e-01 -6.82066381e-01 -4.21927422e-01 4.28392649e-01]
|
[10.166189193725586, 0.7826024889945984]
|
7dd9166b-f4a2-469e-b93d-d44e8cb7efa3
|
collaborative-agent-gameplay-in-the-pandemic
|
2103.11388
| null |
https://arxiv.org/abs/2103.11388v1
|
https://arxiv.org/pdf/2103.11388v1.pdf
|
Collaborative Agent Gameplay in the Pandemic Board Game
|
While artificial intelligence has been applied to control players' decisions in board games for over half a century, little attention is given to games with no player competition. Pandemic is an exemplar collaborative board game where all players coordinate to overcome challenges posed by events occurring during the game's progression. This paper proposes an artificial agent which controls all players' actions and balances chances of winning versus risk of losing in this highly stochastic environment. The agent applies a Rolling Horizon Evolutionary Algorithm on an abstraction of the game-state that lowers the branching factor and simulates the game's stochasticity. Results show that the proposed algorithm can find winning strategies more consistently in different games of varying difficulty. The impact of a number of state evaluation metrics is explored, balancing between optimistic strategies that favor winning and pessimistic strategies that guard against losing.
|
['Antonios Liapis', 'Konstantinos Sfikas']
|
2021-03-21
| null | null | null | null |
['board-games']
|
['playing-games']
|
[-4.78406921e-02 3.75012100e-01 8.98216963e-02 2.07117960e-01
6.64928481e-02 -6.29807115e-01 4.57382858e-01 3.16970080e-01
-8.08688641e-01 1.00958955e+00 -2.09068507e-02 -4.37331885e-01
-5.26114345e-01 -1.14463603e+00 2.28109777e-01 -4.94121730e-01
-1.48891658e-01 7.47944713e-01 7.34905899e-01 -8.51576447e-01
5.21453142e-01 1.65778786e-01 -1.57927501e+00 -4.83956747e-02
5.67472756e-01 6.09576046e-01 1.15771353e-01 1.01136780e+00
4.08685617e-02 1.17606330e+00 -1.00080311e+00 -6.16781533e-01
6.39937341e-01 -5.93722761e-01 -5.82532644e-01 -3.04642886e-01
-9.77303863e-01 -1.24915661e-02 -8.05637911e-02 9.26393569e-01
5.95277369e-01 2.52402484e-01 5.34045458e-01 -1.31575716e+00
3.64577651e-01 6.92020535e-01 -5.31718791e-01 5.71472228e-01
6.29080176e-01 4.10113066e-01 7.15314150e-01 3.91429156e-01
5.83483636e-01 8.42688799e-01 5.10813653e-01 5.77762246e-01
-8.65214467e-01 -5.16423643e-01 7.93872550e-02 2.63529509e-01
-1.25873244e+00 1.79695860e-01 4.84436691e-01 -3.27924371e-01
1.15866232e+00 6.33313596e-01 1.32116377e+00 4.62689757e-01
8.16980660e-01 2.75114149e-01 1.14576757e+00 -5.87857008e-01
7.12196112e-01 -7.45352209e-02 -5.17339289e-01 7.49979690e-02
3.97800803e-01 5.45950890e-01 -3.08944941e-01 -2.84708172e-01
4.53233778e-01 -4.95627761e-01 1.30511358e-01 -2.15876788e-01
-3.23393017e-01 8.40291440e-01 -3.09089154e-01 1.47303939e-01
-9.51156795e-01 -9.13441703e-02 3.45379770e-01 4.15248275e-01
3.11484747e-02 9.63606477e-01 -3.71756107e-01 -8.13047171e-01
-9.09930527e-01 7.45273352e-01 1.03756821e+00 3.11195880e-01
1.20410696e-01 -1.60471983e-02 -2.68233567e-01 2.66905665e-01
1.94681734e-01 -1.94276318e-01 3.44569743e-01 -9.20496225e-01
7.20817372e-02 8.19729149e-01 3.79912406e-01 -1.06780887e+00
-5.26361585e-01 -3.86520803e-01 -2.70147800e-01 8.34706664e-01
4.20308828e-01 -5.27372241e-01 -3.80253255e-01 1.53837347e+00
3.76641214e-01 -1.02719218e-01 -5.54099912e-03 5.24922132e-01
3.67264524e-02 6.05994463e-01 6.67377235e-03 -5.55903733e-01
1.14226770e+00 -3.50325882e-01 -5.96354246e-01 -1.73006535e-01
9.80428979e-02 -4.83330876e-01 3.66543472e-01 6.82829499e-01
-1.68054008e+00 1.14107795e-01 -9.68514800e-01 8.39073718e-01
-2.01264054e-01 -7.15122342e-01 6.16723955e-01 1.18448174e+00
-1.08124375e+00 4.17301506e-01 -7.25281954e-01 -3.20280820e-01
-1.14025995e-02 5.69611311e-01 4.65149164e-01 6.08060420e-01
-1.39752722e+00 1.26096308e+00 6.37096763e-01 2.34364867e-02
-1.01886010e+00 -3.45380068e-01 -2.72207767e-01 2.93443173e-01
8.74990880e-01 -5.76501310e-01 1.29007387e+00 -8.10631752e-01
-1.95707130e+00 7.09841669e-01 7.31905937e-01 -6.41705334e-01
1.10613668e+00 4.89091724e-01 -1.52711049e-01 -3.87190104e-01
-1.24511138e-01 2.09872693e-01 3.86666656e-01 -7.42094159e-01
-1.00740123e+00 -2.59935677e-01 6.35923684e-01 8.49019289e-01
6.09946884e-02 3.58060896e-01 -8.77074450e-02 -3.01486313e-01
-3.31244439e-01 -9.22220469e-01 -1.01257598e+00 -9.32163119e-01
-1.66670084e-01 -4.19545174e-02 -5.18857911e-02 -1.73764244e-01
1.83497131e+00 -1.49475539e+00 2.87492245e-01 3.10329795e-01
9.01870895e-03 1.78911880e-01 3.05641860e-01 1.02935922e+00
2.94408172e-01 1.54196396e-01 2.39058316e-01 3.66295367e-01
4.84950133e-02 2.12643042e-01 5.98852243e-03 6.18967675e-02
-3.02410036e-01 2.43430018e-01 -6.07384801e-01 -1.11218289e-01
6.55115535e-03 -1.98519424e-01 -8.06583881e-01 2.84197539e-01
-2.22181275e-01 4.99921963e-02 -7.01336980e-01 2.58024573e-01
4.69649434e-01 4.03811097e-01 6.03723645e-01 8.30778003e-01
-6.92054033e-01 2.07387596e-01 -1.57998323e+00 8.80065084e-01
-1.11915469e-01 1.77522615e-01 2.35572353e-01 -8.01313281e-01
8.08420718e-01 2.52684236e-01 4.53583598e-01 -5.62104702e-01
6.94561481e-01 -6.74389070e-03 5.71559787e-01 -2.74299204e-01
8.25982988e-01 -3.17630917e-01 -3.69941115e-01 6.18783474e-01
-5.56721568e-01 -5.72618306e-01 4.43459481e-01 2.97776252e-01
1.33413923e+00 3.36002707e-02 8.46324801e-01 -3.02970827e-01
4.32501644e-01 4.42122936e-01 9.37643588e-01 1.10638642e+00
-7.35496521e-01 1.34999696e-02 9.78052080e-01 -5.18821418e-01
-8.25354040e-01 -7.31697083e-01 5.85907519e-01 1.05828929e+00
4.87707317e-01 -4.59314018e-01 -8.57790768e-01 -2.22736597e-03
-2.47735724e-01 8.78615975e-01 -8.24509382e-01 -4.31401610e-01
-4.35005784e-01 -8.77085149e-01 4.10541862e-01 -1.32780209e-01
3.83705020e-01 -1.26741433e+00 -1.62684929e+00 5.47659397e-01
1.76592961e-01 -3.67956430e-01 -1.87026516e-01 1.93368137e-01
-4.86374408e-01 -1.09954572e+00 -1.86339766e-01 -4.99552973e-02
7.50004426e-02 -1.46691829e-01 1.20656312e+00 1.27460986e-01
-5.16411126e-01 2.63784230e-01 -3.14944744e-01 -9.98820066e-01
-6.24208212e-01 -2.54431199e-02 -5.61666824e-02 -4.82816458e-01
3.32308799e-01 -3.79774660e-01 -6.94918096e-01 3.48817199e-01
-8.11759293e-01 1.28952429e-01 8.94452184e-02 6.44623518e-01
1.51898220e-01 7.76449919e-01 4.33835745e-01 -5.04203677e-01
1.43243527e+00 -5.40293753e-01 -8.45609546e-01 2.44456872e-01
-4.00689095e-01 -1.15601726e-01 3.84505004e-01 -4.90167290e-01
-1.11991012e+00 -2.58403093e-01 3.13498437e-01 2.32041657e-01
2.78717101e-01 4.26635653e-01 -8.90104622e-02 2.30525106e-01
6.43066585e-01 7.07813725e-02 1.03631742e-01 1.50300771e-01
-8.97586495e-02 4.06285554e-01 4.77686115e-02 -6.95185423e-01
4.14799750e-01 2.23398089e-01 -7.01088384e-02 -1.71556875e-01
-1.01961464e-01 -1.61422014e-01 -4.94309403e-02 -9.09590960e-01
5.94486415e-01 -4.89325464e-01 -1.30673075e+00 6.00658298e-01
-6.17847323e-01 -2.24074438e-01 -5.13317406e-01 1.78882301e-01
-7.14290559e-01 -1.19099744e-01 -2.55914211e-01 -1.46495378e+00
-5.68448491e-02 -9.75288391e-01 4.57227640e-02 9.81485665e-01
-4.50385958e-01 -6.85548961e-01 7.28434145e-01 2.84410805e-01
5.46759427e-01 4.62261558e-01 3.36348116e-01 -4.78518933e-01
-4.23341423e-01 -3.11080098e-01 6.12117410e-01 -3.05994809e-01
-1.25514269e-01 3.27969611e-01 -2.09791109e-01 -2.74243951e-01
1.27826244e-01 -3.27539928e-02 6.38910607e-02 5.60232937e-01
2.68464386e-01 -3.75295699e-01 -3.10536027e-01 2.53690388e-02
1.43448234e+00 1.10189879e+00 8.99687588e-01 1.16374910e+00
-3.33259761e-01 8.19518089e-01 8.91769767e-01 1.08331037e+00
3.74230415e-01 7.19362319e-01 8.46246958e-01 4.53764319e-01
6.04624391e-01 6.80507952e-03 3.31432194e-01 1.67929113e-01
-5.98658264e-01 -4.08911407e-01 -9.75348055e-01 5.10098696e-01
-1.84729600e+00 -1.35437202e+00 2.22637549e-01 2.28127503e+00
8.09152544e-01 8.66356850e-01 6.66399717e-01 2.57217854e-01
9.51873600e-01 -1.05565088e-02 -5.08378863e-01 -1.20583546e+00
-4.61217389e-02 2.58456081e-01 6.92656934e-01 5.84867895e-01
-5.15301347e-01 1.04577470e+00 6.75650406e+00 8.79762888e-01
-7.58635640e-01 -1.44400284e-01 8.12262952e-01 -8.15787435e-01
-1.82629660e-01 3.25370520e-01 -4.65043366e-01 6.29515231e-01
8.34998131e-01 -8.58700991e-01 5.70904315e-01 5.40320635e-01
7.24336088e-01 -7.15748847e-01 -2.19367981e-01 2.56128997e-01
-4.25963849e-01 -1.32186067e+00 -4.06741619e-01 2.91502982e-01
8.43410611e-01 -3.81405294e-01 -3.37995254e-02 3.53599936e-01
1.29805970e+00 -1.11988854e+00 1.08720589e+00 5.48869550e-01
2.23391280e-01 -1.44471657e+00 7.48977780e-01 7.31638491e-01
-9.00648892e-01 -6.23680294e-01 -1.06813870e-01 -1.05362642e+00
3.69660527e-01 -4.74939123e-02 -8.87368381e-01 3.82844299e-01
6.60226226e-01 -5.45865536e-01 -7.34807327e-02 1.17783225e+00
-1.30378857e-01 2.96437472e-01 -4.40757066e-01 -4.51622576e-01
6.26449585e-01 -6.01806998e-01 7.84839988e-01 4.67041105e-01
3.22312981e-01 7.71457136e-01 6.28931895e-02 4.57870990e-01
8.45618248e-01 5.22398949e-02 -4.21588629e-01 1.85845997e-02
7.58082628e-01 8.38567972e-01 -1.08678269e+00 9.12729558e-03
3.00017387e-01 5.82145393e-01 -1.30767494e-01 -1.13786265e-01
-8.87071013e-01 -2.97892183e-01 8.36266518e-01 3.24680537e-01
1.13994926e-01 1.51655361e-01 -5.79964876e-01 -5.42061865e-01
-4.90579277e-01 -1.13700187e+00 6.69856846e-01 -6.72717750e-01
-5.74167788e-01 7.57781684e-01 1.49408370e-01 -7.83479095e-01
-5.75762749e-01 -1.77090496e-01 -1.25111377e+00 9.55809236e-01
-7.38277018e-01 -2.50118017e-01 2.94435769e-01 3.09666365e-01
3.59390914e-01 -4.27869380e-01 3.74194652e-01 -4.14955318e-01
-8.89514446e-01 2.69990832e-01 6.71677059e-03 -5.94655991e-01
-6.40528277e-02 -1.05543160e+00 8.47849473e-02 8.51309776e-01
-5.53522289e-01 3.45826179e-01 1.33932388e+00 -8.14783216e-01
-9.05429721e-01 -1.44897342e-01 5.46406448e-01 -2.06728980e-01
6.18613660e-01 1.93092003e-02 -2.43479058e-01 -2.00514421e-02
3.56739312e-01 -8.57753277e-01 4.90711540e-01 -1.22061796e-01
3.53711993e-01 1.01790905e-01 -1.44665670e+00 1.11670542e+00
7.83951759e-01 4.09496427e-02 -4.23471212e-01 -1.83724523e-01
5.25560267e-02 -4.67050821e-01 -3.25513065e-01 2.02869698e-01
4.78259057e-01 -1.41880858e+00 6.70944512e-01 -7.35179901e-01
4.18034941e-02 -1.76508874e-01 1.76699609e-01 -1.44358289e+00
-3.66012424e-01 -1.29225659e+00 4.46597308e-01 6.94044352e-01
2.13137507e-01 -4.04024720e-01 1.26230252e+00 7.50963271e-01
3.18707526e-01 -9.28469956e-01 -1.11879194e+00 -6.46320224e-01
2.17450559e-01 -2.44085446e-01 6.70648336e-01 5.93799591e-01
4.28785920e-01 -1.90500140e-01 -3.71341079e-01 -7.85738528e-02
3.76460552e-01 -2.16728821e-01 5.51429033e-01 -1.11299276e+00
-6.58618271e-01 -9.56153512e-01 -6.12348258e-01 -1.14718573e-02
-4.78185713e-01 -1.97979938e-02 -6.15221746e-02 -1.33025849e+00
2.18064517e-01 -3.22615415e-01 -2.54785478e-01 8.56052488e-02
-1.12144887e-01 -1.11078471e-01 5.06545663e-01 -6.34040833e-02
-6.97211802e-01 1.27438307e-01 1.01068377e+00 1.94071472e-01
-5.64513028e-01 2.05025554e-01 -8.72044086e-01 7.08863318e-01
9.74786818e-01 -5.12364805e-01 -6.05626166e-01 3.54363948e-01
8.81454170e-01 5.90395629e-01 -3.86721492e-01 -1.00149679e+00
4.16500270e-01 -8.73446465e-01 -4.32411015e-01 -2.70023972e-01
-1.10103590e-02 -4.07236934e-01 9.63707089e-01 1.14912879e+00
-2.06025094e-01 4.54748839e-01 3.31105173e-01 5.51948905e-01
6.25611171e-02 -4.94495958e-01 8.31799746e-01 -3.10471147e-01
-5.06356001e-01 -1.16651259e-01 -1.25872874e+00 3.87254469e-02
1.85773826e+00 -7.72923946e-01 1.76468581e-01 -7.02675402e-01
-7.84301102e-01 5.38339317e-01 5.65218508e-01 2.12978646e-01
-4.92448583e-02 -6.67681694e-01 -7.86541998e-01 -3.27184826e-01
-4.95847285e-01 -5.38578153e-01 4.39941645e-01 3.61637712e-01
-1.17064178e+00 2.12560982e-01 -7.81632185e-01 1.36724606e-01
-1.46352208e+00 3.49913657e-01 9.50243533e-01 -1.11289370e+00
-1.70198649e-01 1.04258239e+00 -1.07042663e-01 -1.46169946e-01
9.10443887e-02 3.72113407e-01 -5.46899259e-01 3.50833714e-01
5.66838503e-01 7.68133640e-01 -1.29020691e-01 -3.83855671e-01
-5.03349125e-01 -3.29138972e-02 -8.10483173e-02 -6.26301646e-01
1.23383892e+00 -2.05808356e-01 -4.54378314e-02 1.33663565e-01
-2.28745490e-01 -3.73314069e-05 -1.52283514e+00 3.30695122e-01
8.20569843e-02 -4.96563047e-01 5.98518141e-02 -1.12507927e+00
-7.95416594e-01 2.45932877e-01 3.44926268e-01 8.30825984e-01
1.40219223e+00 -7.33692706e-01 2.27709472e-01 -1.21862620e-01
7.85222769e-01 -1.46982944e+00 -5.18768691e-02 7.21596777e-01
4.67633635e-01 -5.02354383e-01 3.14124152e-02 -3.30526568e-02
-9.80151296e-01 7.09799528e-01 7.01103926e-01 -2.66550064e-01
2.91547656e-01 6.40836179e-01 7.09424466e-02 -2.14210853e-01
-1.38785660e+00 -1.85215503e-01 -5.90590060e-01 7.00578392e-01
-6.24965169e-02 2.71144450e-01 -1.03040206e+00 7.20096886e-01
-4.24149215e-01 5.54963537e-02 1.26895547e+00 1.12907732e+00
-6.14450157e-01 -1.27424979e+00 -6.96766078e-01 3.61741215e-01
-7.75380552e-01 1.14134744e-01 -7.17112780e-01 9.17614639e-01
5.14259696e-01 1.12980270e+00 1.32880867e-01 -2.06846371e-01
2.19545275e-01 -2.82911986e-01 3.21004570e-01 -5.61424971e-01
-1.33717144e+00 -4.17724699e-02 2.23666728e-01 -4.31651920e-01
-2.84395576e-03 -9.67743874e-01 -9.89911258e-01 -7.41505146e-01
-2.60697007e-01 6.90915108e-01 4.84258473e-01 6.79407001e-01
1.29857093e-01 7.34898984e-01 5.31033278e-01 -5.21892548e-01
-7.60467529e-01 -4.99577582e-01 -7.68299401e-01 -2.51405299e-01
-4.06998277e-01 -8.01004171e-01 -1.32291749e-01 -7.54856646e-01]
|
[3.4585366249084473, 1.5494972467422485]
|
670eb1e6-287b-4aa4-a904-3f00d48cee13
|
alexsis-pt-a-new-resource-for-portuguese
|
2209.09034
| null |
https://arxiv.org/abs/2209.09034v1
|
https://arxiv.org/pdf/2209.09034v1.pdf
|
ALEXSIS-PT: A New Resource for Portuguese Lexical Simplification
|
Lexical simplification (LS) is the task of automatically replacing complex words for easier ones making texts more accessible to various target populations (e.g. individuals with low literacy, individuals with learning disabilities, second language learners). To train and test models, LS systems usually require corpora that feature complex words in context along with their candidate substitutions. To continue improving the performance of LS systems we introduce ALEXSIS-PT, a novel multi-candidate dataset for Brazilian Portuguese LS containing 9,605 candidate substitutions for 387 complex words. ALEXSIS-PT has been compiled following the ALEXSIS protocol for Spanish opening exciting new avenues for cross-lingual models. ALEXSIS-PT is the first LS multi-candidate dataset that contains Brazilian newspaper articles. We evaluated four models for substitute generation on this dataset, namely mDistilBERT, mBERT, XLM-R, and BERTimbau. BERTimbau achieved the highest performance across all evaluation metrics.
|
['Tharindu Ranasinghe', 'Marcos Zampieri', 'Kai North']
|
2022-09-19
| null |
https://aclanthology.org/2022.coling-1.529
|
https://aclanthology.org/2022.coling-1.529.pdf
|
coling-2022-10
|
['lexical-simplification', 'xlm-r']
|
['natural-language-processing', 'natural-language-processing']
|
[-9.00574308e-03 3.90895218e-01 -3.80932748e-01 -1.64849356e-01
-1.12041259e+00 -3.66663009e-01 4.60676670e-01 4.20150369e-01
-8.75801504e-01 1.25664628e+00 6.03473485e-01 -2.34790757e-01
3.11118633e-01 -6.27157748e-01 -7.53259778e-01 9.71491113e-02
7.22788632e-01 8.91551197e-01 9.31480229e-02 -8.21750045e-01
3.60108688e-02 4.29081976e-01 -1.36817753e+00 6.52570724e-01
1.52609074e+00 1.00965418e-01 5.00078738e-01 4.46934998e-01
-2.77048528e-01 5.24615169e-01 -9.09048021e-01 -1.09786808e+00
1.89096946e-02 -3.95028353e-01 -6.45605087e-01 -6.30773902e-01
7.19144523e-01 2.16659814e-01 1.66372791e-01 1.00214994e+00
8.23365271e-01 4.38879691e-02 7.01032519e-01 -7.61464059e-01
-9.96635497e-01 1.54654527e+00 -1.40613496e-01 2.52352953e-01
7.65526056e-01 -1.47675211e-02 1.12754858e+00 -1.40364277e+00
9.72378612e-01 1.78033900e+00 9.84811664e-01 7.00410187e-01
-1.17216933e+00 -6.33620679e-01 1.73093930e-01 4.95393813e-01
-1.37146342e+00 -6.63006365e-01 3.01182628e-01 -2.78220195e-02
1.69180083e+00 5.77490568e-01 9.23764646e-01 1.34611356e+00
-2.90865507e-02 1.01967466e+00 1.10875511e+00 -1.06542170e+00
-3.64467829e-01 1.82015210e-01 1.79264069e-01 5.16716659e-01
5.96064091e-01 -2.41898730e-01 -7.12755501e-01 1.60570621e-01
3.85504439e-02 -6.30597413e-01 -2.32606575e-01 4.98659819e-01
-1.30789232e+00 8.77945125e-01 -3.99646815e-03 4.07432407e-01
-2.07027704e-01 -2.58138865e-01 5.27417600e-01 5.07282197e-01
3.47483695e-01 1.02235425e+00 -8.77507329e-01 -2.42162719e-01
-6.75730884e-01 5.16106844e-01 8.92864883e-01 1.47029018e+00
2.79721439e-01 2.43828192e-01 -4.42480981e-01 1.24387145e+00
2.01118249e-03 7.98386693e-01 9.04787838e-01 -6.04362369e-01
8.53661120e-01 5.56193113e-01 -1.07737847e-01 -3.56252611e-01
-5.33609748e-01 -2.59462833e-01 -4.20612216e-01 -4.41462338e-01
2.18351796e-01 -1.44530833e-01 -7.49778509e-01 1.75423360e+00
9.13084000e-02 -4.88101304e-01 2.51662612e-01 2.48203762e-02
1.31277728e+00 6.52269185e-01 3.03898960e-01 -4.36950356e-01
1.26913548e+00 -1.26603365e+00 -8.04560304e-01 -3.33454967e-01
1.02479589e+00 -1.07677174e+00 1.93160319e+00 5.23625135e-01
-1.39483380e+00 -3.86018962e-01 -8.92834365e-01 -6.83027744e-01
-7.92816639e-01 4.92716312e-01 3.19490075e-01 7.85580277e-01
-7.88739800e-01 3.92105490e-01 -4.10655469e-01 -4.02851194e-01
2.56853193e-01 2.05228820e-01 -4.45290804e-01 -1.84570000e-01
-1.41401768e+00 1.57704842e+00 6.28986537e-01 -5.22491515e-01
-3.44480604e-01 -1.11776316e+00 -1.06282198e+00 -5.01070432e-02
1.59317777e-01 -5.43390930e-01 1.54104316e+00 -9.75411773e-01
-1.50682259e+00 1.10441017e+00 -3.25623393e-01 -4.75037485e-01
6.17864013e-01 -5.46160340e-01 -9.04657006e-01 -3.49944293e-01
6.16801143e-01 7.65638351e-01 5.73807657e-01 -8.05299878e-01
-8.33817065e-01 5.70999160e-02 -1.28268749e-01 4.14918572e-01
-2.70447493e-01 3.50540489e-01 -7.81773552e-02 -1.18646502e+00
-3.26335520e-01 -8.76736343e-01 5.50105684e-02 -6.30053878e-01
-6.22068405e-01 -5.81182122e-01 4.13547844e-01 -1.02914321e+00
1.57067823e+00 -1.62645042e+00 2.19619825e-01 -1.57541245e-01
-2.53836840e-01 8.67756665e-01 -3.12086821e-01 5.20477593e-01
-1.88643858e-01 4.48577404e-01 -1.11557446e-01 -6.67119741e-01
1.25721827e-01 3.23241830e-01 5.28276525e-02 -4.57163528e-02
2.33458251e-01 1.22525060e+00 -9.27969575e-01 -4.49935257e-01
1.50727313e-02 1.57241836e-01 -8.59462619e-01 -2.57416874e-01
-4.07318592e-01 1.60915047e-01 1.22334369e-01 7.29943216e-01
3.32958698e-01 4.17070210e-01 -3.18146832e-02 -5.00372574e-02
-3.12267274e-01 8.95363569e-01 -7.86422431e-01 1.49047983e+00
-8.54639173e-01 5.98272860e-01 -4.83541697e-01 -3.74331355e-01
4.76559132e-01 2.27472737e-01 -1.68866012e-02 -9.81081784e-01
6.66961148e-02 8.36292684e-01 6.60463870e-02 -4.52263594e-01
9.35060620e-01 -8.98852497e-02 -3.56867999e-01 6.84835464e-02
-3.28664370e-02 -6.24712169e-01 9.75596607e-01 9.33026597e-02
9.64155436e-01 -5.29422015e-02 8.75449240e-01 -4.42208081e-01
8.36267829e-01 1.75924137e-01 3.97244006e-01 6.09258890e-01
1.92144766e-01 4.83677894e-01 1.44721031e-01 -1.53728724e-01
-1.06450391e+00 -9.96012270e-01 -1.43534094e-01 1.25231314e+00
-4.53482687e-01 -7.45006621e-01 -7.15120912e-01 -7.26665437e-01
1.39439879e-02 1.57300282e+00 -3.04172128e-01 -2.45871946e-01
-1.11187363e+00 -4.62263167e-01 8.59536707e-01 1.09555824e-02
2.89081186e-01 -1.55227256e+00 -1.56270295e-01 4.05333608e-01
-4.73096400e-01 -1.10938752e+00 -6.00808620e-01 -4.76763323e-02
-3.25398356e-01 -8.80031645e-01 -5.44333994e-01 -1.17589068e+00
3.73585582e-01 -1.17866300e-01 1.67951369e+00 4.97863330e-02
-1.46233410e-01 -6.01740666e-02 -5.68485677e-01 -1.04025280e+00
-9.41723406e-01 5.92504561e-01 1.15488164e-01 -9.34799969e-01
6.03350163e-01 -2.26428390e-01 2.30652124e-01 -3.57807875e-01
-6.01526499e-01 1.13403618e-01 5.68530977e-01 9.86308396e-01
8.28972936e-01 -4.95509714e-01 7.36293674e-01 -1.50457573e+00
8.76954854e-01 -4.66662735e-01 -2.58477092e-01 4.24891293e-01
-5.03238261e-01 -9.24891084e-02 9.80963767e-01 -6.00411117e-01
-1.15264630e+00 -2.49760360e-01 -7.71374524e-01 3.35706264e-01
1.11633435e-01 6.82252705e-01 -3.99825901e-01 2.53865898e-01
1.03726697e+00 -1.29415646e-01 -5.35221994e-01 -7.39272416e-01
6.99285865e-01 6.41898870e-01 6.14091277e-01 -6.12595856e-01
3.78769159e-01 -1.56585798e-01 -4.16269839e-01 -1.15800822e+00
-1.07963669e+00 -4.67813350e-02 -5.17736793e-01 2.94814736e-01
5.24475157e-01 -9.56093729e-01 -1.57172993e-01 4.32391226e-01
-1.38913763e+00 -3.21323395e-01 -7.44718492e-01 4.01101261e-01
-2.63792634e-01 8.37843791e-02 -4.90125924e-01 -9.21084210e-02
-5.98976135e-01 -1.04243863e+00 1.02329683e+00 6.76110908e-02
-9.39796507e-01 -9.67228353e-01 1.35878906e-01 2.30399907e-01
2.49505267e-01 -3.91817763e-02 1.49074316e+00 -7.92368472e-01
1.05159298e-01 -1.63045779e-01 1.16670981e-01 5.51426232e-01
1.15528636e-01 1.07062452e-01 -4.63267654e-01 -1.35010421e-01
-4.18577373e-01 -4.41883594e-01 6.11437142e-01 3.05915058e-01
8.57088625e-01 -5.07389128e-01 -4.69107777e-02 5.33124745e-01
1.01862454e+00 1.53046027e-02 5.49454391e-01 5.27150512e-01
9.53569055e-01 3.18322808e-01 5.22772670e-01 1.46500871e-01
8.10683966e-01 6.32747591e-01 -1.07175156e-01 2.32982892e-03
-7.79640019e-01 -3.87834460e-01 7.56226540e-01 1.59151101e+00
2.78613895e-01 -3.79365325e-01 -8.98770511e-01 7.36048281e-01
-1.35066080e+00 -7.87198007e-01 -5.53640962e-01 1.92848253e+00
1.48553276e+00 6.59077913e-02 -7.97907263e-02 2.54530609e-01
5.39327323e-01 -1.75251633e-01 -2.91561127e-01 -8.89330864e-01
-7.60056317e-01 8.69682372e-01 2.27492064e-01 7.41974950e-01
-7.58303523e-01 1.71244037e+00 6.15917110e+00 1.23997152e+00
-8.85414839e-01 3.33706111e-01 2.02368543e-01 -2.11143732e-01
-6.84171557e-01 -1.34170800e-01 -1.31368792e+00 3.98280263e-01
1.01050067e+00 -4.81832117e-01 4.86479938e-01 5.73924661e-01
1.19783469e-01 -6.35965392e-02 -1.20208943e+00 8.74836743e-01
3.75596225e-01 -1.21227396e+00 4.82366890e-01 -6.76815569e-01
1.18095338e+00 2.69257843e-01 -2.50269473e-02 9.51674640e-01
4.88212675e-01 -1.20246780e+00 1.12477946e+00 3.66619855e-01
1.01275933e+00 -8.53205442e-01 5.04334033e-01 3.26682538e-01
-7.95413792e-01 1.43944114e-01 -3.23181868e-01 -7.40446988e-03
1.23293258e-01 3.86788994e-01 -8.85452747e-01 2.68538535e-01
4.74858105e-01 8.09041440e-01 -1.05671096e+00 7.44656682e-01
-7.18625903e-01 5.61396182e-01 -3.31378609e-01 -3.92926872e-01
-3.44770961e-02 3.11715435e-02 8.68615746e-01 1.65264785e+00
5.07638037e-01 -2.40258485e-01 -5.33466600e-02 3.91475677e-01
-3.14312696e-01 8.84237409e-01 -4.83283669e-01 -7.61906011e-03
9.63136256e-01 8.93618226e-01 -2.00186327e-01 -5.69733918e-01
-5.20540595e-01 9.19875622e-01 6.73655510e-01 1.88635588e-02
-6.31602705e-01 -4.69399989e-01 4.68691885e-01 1.18273003e-02
-7.01669976e-02 1.52871758e-02 -5.58490813e-01 -1.19225538e+00
-8.46251007e-03 -1.64402878e+00 3.02031159e-01 -6.15152478e-01
-1.33186591e+00 6.81361198e-01 1.62417054e-01 -7.81751037e-01
-1.23699248e-01 -6.06085777e-01 -2.39980876e-01 9.85753477e-01
-1.57377768e+00 -1.40133119e+00 3.73789929e-02 5.90963066e-01
1.08673847e+00 -5.76417744e-01 9.97991323e-01 5.67607880e-01
-6.37617350e-01 1.03775680e+00 7.43050128e-02 -3.60690087e-01
7.89875686e-01 -1.34119642e+00 7.52329230e-01 9.49161351e-01
2.19270959e-01 5.00605106e-01 6.48237884e-01 -8.81012261e-01
-8.18679512e-01 -1.23608732e+00 1.71178448e+00 -5.32292068e-01
6.32442057e-01 -2.20174655e-01 -7.69429624e-01 8.24425757e-01
3.69784474e-01 -6.39652908e-01 4.71271247e-01 -1.77430853e-01
-1.06831193e-01 1.89254686e-01 -1.16250134e+00 1.25409770e+00
1.37983656e+00 -3.84263188e-01 -8.10334027e-01 7.42939413e-01
1.00493240e+00 -7.31692135e-01 -7.46747613e-01 2.98459291e-01
2.80958325e-01 -5.88849187e-01 6.84482396e-01 -8.14593136e-01
4.45024908e-01 1.82890072e-01 -1.84621483e-01 -1.90736520e+00
-1.58738330e-01 -8.76200080e-01 -1.98712647e-02 1.24231601e+00
1.02341020e+00 -6.23370528e-01 2.53993958e-01 2.05241218e-01
-7.81256855e-01 -5.35122335e-01 -1.12084591e+00 -7.90569663e-01
6.45120025e-01 -4.90717202e-01 9.91778255e-01 9.54182088e-01
-1.89678267e-01 6.22949183e-01 4.18895595e-02 -2.40386724e-01
1.20645501e-02 -5.33800840e-01 7.47253120e-01 -9.90106344e-01
-1.20492309e-01 -7.52348363e-01 1.30258903e-01 -6.64506733e-01
6.65176570e-01 -1.45527112e+00 -2.88112164e-01 -1.59085584e+00
-2.17806116e-01 -5.92864931e-01 4.38132107e-01 6.64234102e-01
-4.89432275e-01 1.68396622e-01 1.50996402e-01 -1.64199278e-01
-2.13879853e-01 6.96464837e-01 1.26810789e+00 -7.56799132e-02
-3.23164016e-01 -4.40529883e-02 -7.57473171e-01 8.59906912e-01
8.27909946e-01 -6.28427029e-01 -2.26918727e-01 -9.28244591e-01
5.09582460e-01 -5.23157239e-01 -3.54811877e-01 -7.80281842e-01
-1.85810834e-01 -1.80458978e-01 1.13685451e-01 -5.30937195e-01
2.36616507e-01 -4.72478449e-01 1.52688295e-01 5.93109190e-01
-3.56267959e-01 7.61853695e-01 2.79522330e-01 -4.22031373e-01
-4.77812178e-02 -6.99800611e-01 8.76875043e-01 -3.38339627e-01
-5.29467881e-01 -7.38280937e-02 -5.12164056e-01 7.39736497e-01
8.66584301e-01 3.22236605e-02 -6.16493642e-01 -1.74250361e-02
-5.98296046e-01 5.51198833e-02 2.38018706e-01 6.46452188e-01
4.64242995e-01 -1.41634786e+00 -1.09267855e+00 3.83216470e-01
2.87207782e-01 6.39933124e-02 -3.58965814e-01 5.96143425e-01
-1.02507567e+00 4.91610140e-01 -1.17643259e-01 2.32274234e-02
-1.28789783e+00 1.55944988e-01 1.51685774e-01 -4.18084770e-01
-4.27901089e-01 1.16569757e+00 -5.22973061e-01 -8.69485140e-01
-3.05663678e-04 -7.09535837e-01 -3.60507578e-01 2.20554233e-01
3.70277643e-01 8.71575892e-01 4.96668696e-01 -1.04515553e+00
-7.16399327e-02 9.74490047e-02 -3.18193883e-01 -1.58847332e-01
1.19523656e+00 -9.05156508e-02 -4.19588536e-01 3.59679163e-01
9.18359995e-01 8.83437455e-01 -1.42892614e-01 -2.68761277e-01
4.74044949e-01 -1.43228531e-01 -3.04717809e-01 -1.02711344e+00
-6.38226211e-01 4.58566636e-01 3.93975340e-02 -2.04282463e-01
7.95685470e-01 -2.57614911e-01 1.02711034e+00 6.03317082e-01
3.37134808e-01 -1.42857611e+00 -9.44295675e-02 1.26137781e+00
1.34271109e+00 -1.14298546e+00 6.69875965e-02 -5.38103461e-01
-7.59802878e-01 8.91953170e-01 7.47752845e-01 2.10249424e-02
5.13371050e-01 -5.20648174e-02 1.23007625e-01 5.21951951e-02
-7.19713688e-01 -2.04474658e-01 4.30019289e-01 7.11130857e-01
8.30758333e-01 5.15191078e-01 -1.02384591e+00 9.41005349e-01
-1.21503162e+00 -5.33870041e-01 6.30822539e-01 6.44393325e-01
-1.78461030e-01 -1.55259693e+00 -1.40816182e-01 7.28756130e-01
-5.38959622e-01 -8.97443473e-01 -6.75054014e-01 1.14751256e+00
4.47785199e-01 8.25843751e-01 -3.05313379e-01 1.58458322e-01
9.83239472e-01 1.69039562e-01 5.78842640e-01 -1.24070585e+00
-1.00037444e+00 -2.67317086e-01 7.63781667e-01 -8.98959786e-02
6.63354397e-02 -9.88797843e-01 -1.28831208e+00 -4.76586580e-01
-2.50363588e-01 -7.90650249e-02 6.04275048e-01 9.40023839e-01
-1.02428822e-02 5.52485466e-01 9.39890742e-02 -7.58815706e-01
-4.61183459e-01 -1.25758183e+00 -1.36734009e-01 3.12057406e-01
2.84218788e-02 -4.21365798e-01 -5.92249678e-03 -1.99371099e-01]
|
[10.934338569641113, 10.416213989257812]
|
2501d1ec-5144-498c-8b89-512860c109da
|
your-attention-deserves-attention-a-self
|
2203.12570
| null |
https://arxiv.org/abs/2203.12570v1
|
https://arxiv.org/pdf/2203.12570v1.pdf
|
Your "Attention" Deserves Attention: A Self-Diversified Multi-Channel Attention for Facial Action Analysis
|
Visual attention has been extensively studied for learning fine-grained features in both facial expression recognition (FER) and Action Unit (AU) detection. A broad range of previous research has explored how to use attention modules to localize detailed facial parts (e,g. facial action units), learn discriminative features, and learn inter-class correlation. However, few related works pay attention to the robustness of the attention module itself. Through experiments, we found neural attention maps initialized with different feature maps yield diverse representations when learning to attend the identical Region of Interest (ROI). In other words, similar to general feature learning, the representational quality of attention maps also greatly affects the performance of a model, which means unconstrained attention learning has lots of randomnesses. This uncertainty lets conventional attention learning fall into sub-optimal. In this paper, we propose a compact model to enhance the representational and focusing power of neural attention maps and learn the "inter-attention" correlation for refined attention maps, which we term the "Self-Diversified Multi-Channel Attention Network (SMA-Net)". The proposed method is evaluated on two benchmark databases (BP4D and DISFA) for AU detection and four databases (CK+, MMI, BU-3DFE, and BP4D+) for facial expression recognition. It achieves superior performance compared to the state-of-the-art methods.
|
['Lijun Yin', 'Geran Zhao', 'Huiyuan Yang', 'Zhihua Li', 'Xiaotian Li']
|
2022-03-23
| null | null | null | null |
['action-analysis', 'facial-expression-recognition']
|
['computer-vision', 'computer-vision']
|
[ 1.43857608e-02 -1.71019658e-01 -1.85095802e-01 -4.56544042e-01
-4.32993114e-01 7.95686096e-02 2.31017694e-01 -5.28541088e-01
-1.78202629e-01 5.11397779e-01 1.61232427e-01 4.69639897e-01
2.27214880e-02 -4.65604484e-01 -5.22519290e-01 -8.75874460e-01
-5.85673414e-02 -2.00011536e-01 -1.11996621e-01 -3.80119562e-01
1.38164654e-01 6.97859108e-01 -1.67874658e+00 4.82155174e-01
4.33643460e-01 1.52532625e+00 3.55279632e-02 1.44544527e-01
-1.03829175e-01 8.95532012e-01 -5.98999977e-01 -3.47994536e-01
-1.99359357e-01 -4.80295867e-01 -8.81102979e-01 1.08594716e-01
2.59129375e-01 -2.49922037e-01 -2.98993886e-01 1.18654990e+00
5.68629205e-01 1.46766454e-01 7.67953038e-01 -1.39566255e+00
-1.06428540e+00 6.91849217e-02 -1.15378368e+00 5.65889895e-01
1.83240667e-01 -1.04380459e-01 9.61217761e-01 -1.21305180e+00
3.57817680e-01 1.59892392e+00 4.39582735e-01 7.91411519e-01
-7.71580637e-01 -1.07369757e+00 3.99012655e-01 4.91552651e-01
-1.69979191e+00 -4.63402599e-01 7.78860629e-01 -3.27034920e-01
8.09434533e-01 1.76124975e-01 4.84883487e-01 1.11945534e+00
4.08175826e-01 1.12093472e+00 9.68792140e-01 -2.44125664e-01
-2.39794388e-01 3.06182411e-02 6.45462051e-02 8.61953974e-01
-2.29472920e-01 -8.52904245e-02 -5.63665450e-01 -4.56906818e-02
7.59908020e-01 3.21095847e-02 -4.49263334e-01 1.60557568e-01
-4.23287988e-01 8.31408501e-01 7.12608397e-01 5.13343394e-01
-5.44142425e-01 1.01200432e-01 5.00051975e-01 2.04243645e-01
6.34470820e-01 -8.06418806e-02 -3.19040179e-01 1.60810426e-02
-4.01511878e-01 -1.80975944e-02 9.39051658e-02 6.42207503e-01
9.86980379e-01 2.80388683e-01 -4.87184554e-01 1.10400856e+00
4.79973912e-01 3.37389141e-01 6.59629345e-01 -5.26416123e-01
1.65841907e-01 6.45376921e-01 -2.85914421e-01 -1.24615824e+00
-3.38236958e-01 -2.32666686e-01 -1.06714070e+00 3.01960230e-01
-6.74799755e-02 -3.30430418e-01 -9.19228375e-01 1.84245241e+00
8.76705647e-02 3.43860954e-01 -2.76219733e-02 1.15706396e+00
1.24325836e+00 6.43697619e-01 3.75872821e-01 -2.64645904e-01
1.31833744e+00 -9.37147319e-01 -9.06179309e-01 -8.71448442e-02
2.18532443e-01 -6.27159357e-01 8.37994874e-01 2.14119777e-01
-8.72480154e-01 -9.93941963e-01 -8.77284169e-01 1.02437444e-01
-2.75553048e-01 1.39437333e-01 8.27304661e-01 5.59095383e-01
-8.37736130e-01 3.30531657e-01 -3.92096788e-01 -2.37309769e-01
1.10795414e+00 5.48735380e-01 -5.97137332e-01 2.85592880e-02
-1.26609850e+00 7.10992575e-01 2.20249202e-02 3.69280219e-01
-9.82629359e-01 -3.87671947e-01 -7.96713054e-01 2.48366505e-01
6.86806217e-02 -3.09263647e-01 9.44388568e-01 -1.81089675e+00
-1.63965929e+00 8.21365058e-01 -2.26324812e-01 4.43808287e-02
-1.45818174e-01 -3.06524038e-01 -5.80977738e-01 1.75317675e-01
-3.35153118e-02 8.37725461e-01 1.04201365e+00 -1.04694402e+00
-5.17053962e-01 -6.09727800e-01 -7.20976433e-03 5.79451770e-02
-4.98100936e-01 6.29358351e-01 -3.87894213e-01 -6.63894236e-01
-1.99117765e-01 -5.70924997e-01 3.23519558e-02 7.30084479e-02
-5.04291914e-02 -6.03040874e-01 1.09262919e+00 -4.22040820e-01
1.25474668e+00 -2.37215710e+00 2.90648490e-01 1.79991201e-01
1.98201332e-02 4.86127824e-01 -3.85059357e-01 -1.35504395e-01
-3.99144262e-01 8.32991078e-02 1.15320772e-01 -1.66852951e-01
-2.53252327e-01 2.50867903e-01 -1.10995039e-01 5.19343436e-01
7.41549075e-01 1.13398838e+00 -6.52505457e-01 -6.04673922e-01
2.70089842e-02 6.54448211e-01 -5.63416898e-01 3.21531743e-01
2.81640649e-01 4.56163824e-01 -6.94259942e-01 1.17266250e+00
6.92889631e-01 -1.48418367e-01 -2.09812999e-01 -3.98860604e-01
2.14377254e-01 -6.18871212e-01 -1.07352173e+00 1.49292684e+00
-4.01526779e-01 6.69865370e-01 1.54195309e-01 -1.02036989e+00
1.22964394e+00 3.59851420e-01 4.39289123e-01 -8.19511652e-01
6.59208238e-01 -5.76435998e-02 4.18627486e-02 -6.66465878e-01
1.30809873e-01 -2.49731600e-01 1.87586173e-01 2.03923002e-01
5.60731351e-01 6.78259075e-01 -3.17668766e-01 -3.63040775e-01
5.98607123e-01 3.06227747e-02 3.93342048e-01 -3.18518102e-01
9.28461015e-01 -7.92288125e-01 6.77629054e-01 3.41502190e-01
-4.87794459e-01 4.47694957e-01 5.77871203e-01 -4.56335962e-01
-3.52898836e-01 -6.13717973e-01 -3.45338374e-01 1.68284512e+00
1.57030776e-01 -3.02825004e-01 -8.26333046e-01 -7.88864255e-01
6.33776486e-02 1.19045071e-01 -1.22487175e+00 -4.68795121e-01
-2.31604099e-01 -8.80143106e-01 4.31219667e-01 7.75910079e-01
6.73626721e-01 -1.53102839e+00 -3.63369584e-01 -5.40211573e-02
1.09902009e-01 -7.40492225e-01 -4.08299565e-01 2.05122847e-02
-3.80508363e-01 -9.73662496e-01 -1.00447881e+00 -8.12122047e-01
7.18352199e-01 1.54994398e-01 8.49807858e-01 1.58067033e-01
-3.94843370e-01 3.20574373e-01 -4.91026700e-01 -6.74102664e-01
1.75019056e-01 -1.89619616e-01 4.78567183e-02 9.47966456e-01
7.54067838e-01 -4.32373524e-01 -5.01732707e-01 4.87065762e-01
-6.95102096e-01 -4.05457377e-01 8.70212138e-01 1.18142772e+00
6.24105453e-01 -3.78128648e-01 8.82024527e-01 -5.54166257e-01
7.57084548e-01 -6.77165091e-01 -1.94662824e-01 2.83938885e-01
-1.28005281e-01 -5.85064813e-02 4.61380482e-01 -4.17014003e-01
-1.03914237e+00 -7.99416527e-02 -3.16762149e-01 -1.00995421e+00
-1.92121208e-01 3.70615631e-01 -5.57303548e-01 -5.05021453e-01
3.29480171e-01 1.53335869e-01 2.88364887e-01 -4.21697766e-01
6.18549250e-02 8.74463499e-01 3.16174418e-01 -5.88625252e-01
1.88398078e-01 2.64365345e-01 -3.29814777e-02 -7.79432893e-01
-7.84218967e-01 -2.70252883e-01 -5.06667078e-01 -3.68739873e-01
8.89003813e-01 -8.19250703e-01 -7.92907774e-01 6.66662216e-01
-1.09026158e+00 5.27766487e-03 -4.07740213e-02 3.43315721e-01
-4.59283412e-01 1.97828755e-01 -4.91955936e-01 -7.67218769e-01
-4.66078043e-01 -1.23670077e+00 1.28376901e+00 5.71020663e-01
-3.31588835e-02 -6.88360691e-01 -7.96597227e-02 -4.65191714e-02
3.90623242e-01 3.40269893e-01 5.26771724e-01 -2.19508886e-01
-1.77527830e-01 -1.39697539e-02 -6.81104124e-01 5.59946418e-01
2.14498878e-01 2.66175233e-02 -1.40870440e+00 -1.30271807e-01
-2.08489433e-01 -6.19213700e-01 1.02662742e+00 4.43778217e-01
1.66754007e+00 -1.70244694e-01 -3.10097069e-01 8.26269269e-01
1.06865120e+00 2.63084263e-01 9.34442222e-01 4.14489061e-02
5.79825163e-01 5.60232580e-01 8.11048329e-01 6.79278731e-01
-7.09564686e-02 7.60591745e-01 5.27603388e-01 -3.90166759e-01
7.65044317e-02 1.11559868e-01 2.61979461e-01 3.35296690e-01
-5.78609765e-01 -6.02241382e-02 -4.20288712e-01 3.97727281e-01
-1.79247093e+00 -1.06429124e+00 5.35662174e-01 1.53901434e+00
4.86041546e-01 -3.20911825e-01 3.73321027e-02 -1.21471241e-01
6.66362822e-01 3.08556408e-01 -4.99987543e-01 -7.17596173e-01
-2.47603461e-01 6.01358354e-01 1.32630561e-02 1.17505975e-01
-1.21596825e+00 9.73848879e-01 5.64252806e+00 1.15238869e+00
-1.44155395e+00 2.65842497e-01 1.06873071e+00 6.66823611e-02
6.43090829e-02 -6.45977616e-01 -8.50150824e-01 4.46166396e-01
6.84349418e-01 4.67407033e-02 2.00295355e-02 1.05716836e+00
-7.84038007e-02 2.00194836e-01 -7.54407942e-01 1.39825320e+00
3.84068191e-01 -1.05815768e+00 3.12523216e-01 -4.05956656e-02
5.56809545e-01 -1.96721554e-01 2.48316005e-01 5.69169581e-01
-2.31490046e-01 -1.53012383e+00 3.30182672e-01 8.13567102e-01
1.01530910e+00 -1.07911408e+00 1.06511700e+00 -1.82513714e-01
-1.28781867e+00 -2.70644873e-01 -6.67474866e-01 1.27161399e-01
-1.33738786e-01 -1.21623456e-01 -2.49575332e-01 4.02892768e-01
1.04825783e+00 9.56278086e-01 -4.91785824e-01 6.35676503e-01
-9.53376368e-02 3.84400547e-01 9.79078636e-02 -1.89133659e-01
4.49555218e-01 -2.34066062e-02 1.78655684e-01 1.27762187e+00
2.03159258e-01 4.10802990e-01 -1.43987685e-01 7.19001174e-01
-1.85837537e-01 3.62183571e-01 -3.17109823e-01 7.54078850e-02
-2.36147538e-01 1.43499923e+00 -2.84322768e-01 -4.58296053e-02
-6.81949437e-01 9.80866969e-01 5.39783299e-01 3.45170677e-01
-9.47381556e-01 -4.31626529e-01 1.13048470e+00 -7.28605501e-03
6.03835642e-01 2.69533396e-01 2.60927737e-01 -9.15920317e-01
-2.21608639e-01 -7.87584424e-01 5.45568824e-01 -8.20174098e-01
-1.37924063e+00 9.56230283e-01 -1.24583073e-01 -9.75131273e-01
3.88917737e-02 -8.79727066e-01 -7.21671581e-01 9.85085309e-01
-1.63724184e+00 -1.29950321e+00 -6.22072816e-01 1.01554441e+00
5.01496136e-01 -4.99706656e-01 8.87236238e-01 5.29704690e-01
-7.96613991e-01 1.21680737e+00 -2.39270166e-01 4.25709814e-01
6.31674945e-01 -6.53076112e-01 -3.00586551e-01 4.14672464e-01
1.33402064e-01 3.22397053e-01 5.14423922e-02 -1.83583602e-01
-1.19104052e+00 -1.19241285e+00 4.06389207e-01 -1.31443724e-01
4.00797844e-01 -1.15577534e-01 -1.04435754e+00 6.28796101e-01
2.26440623e-01 5.75668812e-01 7.08809793e-01 2.40008280e-01
-2.37950400e-01 -5.21070659e-01 -1.05756879e+00 2.34066069e-01
9.46941078e-01 -4.34261590e-01 -2.60074764e-01 4.36194204e-02
2.17611864e-01 -1.27913266e-01 -8.29224408e-01 6.71305835e-01
7.37565160e-01 -9.38185036e-01 7.58461654e-01 -1.03404725e+00
3.07180166e-01 -1.48389354e-01 -2.81737089e-01 -1.13488281e+00
-6.26628458e-01 -3.88844490e-01 -1.62515238e-01 1.27370727e+00
-3.38446535e-02 -4.30294544e-01 5.43877423e-01 2.46658232e-02
-7.85197020e-02 -1.28087556e+00 -1.02523923e+00 -3.12545359e-01
-1.23066001e-01 -1.79044843e-01 7.17805862e-01 8.63139510e-01
-2.16484070e-01 4.41122413e-01 -5.00997841e-01 9.12854671e-02
1.28123790e-01 1.54260293e-01 5.04814148e-01 -1.10722589e+00
-8.35098177e-02 -7.36439645e-01 -9.02170956e-01 -9.50201571e-01
5.47640860e-01 -7.54307687e-01 -2.19031543e-01 -8.42671871e-01
4.62154448e-01 -2.18547314e-01 -7.91729033e-01 7.45287836e-01
-3.38209391e-01 4.50934231e-01 8.13844800e-02 -4.96826544e-02
-7.68395424e-01 9.89581048e-01 1.25272512e+00 -2.07406849e-01
-2.95772981e-02 -5.20046391e-02 -8.81029665e-01 6.57903373e-01
6.67881668e-01 -5.09332642e-02 -1.90448582e-01 -2.49479219e-01
-3.72622252e-01 -1.22996517e-01 3.55288506e-01 -9.42685604e-01
-1.18312733e-02 -1.84304029e-01 7.97505736e-01 -3.54086220e-01
3.57383609e-01 -6.79137409e-01 -2.92439222e-01 7.18865991e-02
-1.16586804e-01 -2.17690557e-01 5.22006691e-01 3.93884450e-01
-6.42388105e-01 -1.22369766e-01 1.09236097e+00 -1.39777035e-01
-1.23029184e+00 8.33379269e-01 -2.63687283e-01 -2.18272135e-02
1.08000505e+00 -3.21639627e-01 -2.87395995e-02 -2.85341531e-01
-7.77048409e-01 1.82015859e-02 -1.95531100e-01 7.71316946e-01
9.07545090e-01 -1.63813508e+00 -8.36614490e-01 5.49302220e-01
3.22630614e-01 -3.12538177e-01 5.15603542e-01 8.29330206e-01
-5.86110353e-02 2.96816647e-01 -8.46138775e-01 -6.20374143e-01
-1.42591465e+00 5.08287728e-01 6.31451488e-01 1.42478108e-01
-1.26876861e-01 1.28001249e+00 7.35785067e-01 1.07105859e-01
1.69851229e-01 3.91271943e-03 -6.60289168e-01 2.18754023e-01
8.65970314e-01 4.56513613e-02 -1.36468813e-01 -1.14997411e+00
-4.53148276e-01 8.97443295e-01 -2.37136781e-01 6.25402987e-01
1.24722481e+00 -1.89818852e-02 -1.54129758e-01 9.18340683e-02
1.48099053e+00 -3.04671884e-01 -1.22168314e+00 -2.15559989e-01
-3.52624446e-01 -5.74010789e-01 2.07967997e-01 -4.57676530e-01
-1.55878639e+00 1.17197943e+00 1.00793648e+00 -3.41828614e-01
1.46114445e+00 2.21723557e-01 4.47458684e-01 4.21418622e-02
1.89954296e-01 -9.71495390e-01 4.25220758e-01 4.09964770e-01
1.30178809e+00 -1.37516356e+00 -1.73560366e-01 -1.93104669e-01
-8.04360926e-01 1.19549727e+00 1.15754128e+00 -1.59318984e-01
9.45411921e-01 1.29216924e-01 -2.75666881e-02 -3.78401220e-01
-5.34444869e-01 -3.49129289e-01 5.31738341e-01 5.42673290e-01
4.60699916e-01 -1.41613424e-01 7.41357878e-02 1.05219162e+00
2.86189318e-01 8.37335810e-02 -1.47328496e-01 6.32578969e-01
-5.36204100e-01 -5.88182151e-01 -2.77708322e-01 3.51973832e-01
-7.03679800e-01 6.86732307e-02 -3.83187085e-01 9.10861790e-01
4.52337235e-01 5.58321595e-01 3.33969146e-01 -4.35012758e-01
2.29973614e-01 -5.84826656e-02 4.69465494e-01 -3.99904221e-01
-4.69606906e-01 -2.96051465e-02 -1.98028401e-01 -6.99954927e-01
-6.72631383e-01 -4.15078431e-01 -1.04409432e+00 -1.14648916e-01
-2.85738409e-01 1.30640760e-01 7.17637986e-02 7.61441231e-01
5.55429041e-01 5.68681359e-01 8.51361036e-01 -7.70444393e-01
-1.94390267e-01 -1.11499906e+00 -7.48693585e-01 3.72995079e-01
2.26577654e-01 -1.23148942e+00 -8.15608203e-02 -3.09316933e-01]
|
[13.610480308532715, 1.659234881401062]
|
7e8ad3fc-8cb3-462b-9c70-a1f10de83bac
|
verbal-valency-frame-detection-and-selection
| null | null |
https://aclanthology.org/W14-2902
|
https://aclanthology.org/W14-2902.pdf
|
Verbal Valency Frame Detection and Selection in Czech and English
| null |
["Zde{\\v{n}}ka Ure{\\v{s}}ov{\\'a}", 'Jan Haji{\\v{c}}', 'Ond{\\v{r}}ej Du{\\v{s}}ek']
|
2014-06-01
| null | null | null |
ws-2014-6
|
['predicate-detection']
|
['natural-language-processing']
|
[-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01]
|
[-7.397832870483398, 3.8665268421173096]
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.